Grammalecte  Changes On Branch c8714ac3a1d21a06

"""
Grammalecte: compile rules
"""

import re
import traceback
import json

import compile_rules_js_convert as jsconv
import compile_rules_graph as crg


dDEF = {}
lFUNCTIONS = []

aRULESET = set()     # set of rule-ids to check if there is several rules with the same id
nRULEWITHOUTNAME = 0

dJSREGEXES = {}

sWORDLIMITLEFT  = r"(?<![\w.,–-])"   # r"(?<![-.,—])\b"  seems slower
sWORDLIMITRIGHT = r"(?![\w–-])"      # r"\b(?!-—)"       seems slower


def prepareFunction (s):
    "convert simple rule syntax to a string of Python code"
    s = s.replace("__also__", "bCondMemo")
    s = s.replace("__else__", "not bCondMemo")
    s = re.sub(r"isStart *\(\)", 'before("^ *$|, *$")', s)
    s = re.sub(r"isRealStart *\(\)", 'before("^ *$")', s)
    s = re.sub(r"isStart0 *\(\)", 'before0("^ *$|, *$")', s)
    s = re.sub(r"isRealStart0 *\(\)", 'before0("^ *$")', s)
    s = re.sub(r"isEnd *\(\)", 'after("^ *$|^,")', s)
    s = re.sub(r"isRealEnd *\(\)", 'after("^ *$")', s)
    s = re.sub(r"isEnd0 *\(\)", 'after0("^ *$|^,")', s)
    s = re.sub(r"isRealEnd0 *\(\)", 'after0("^ *$")', s)
    s = re.sub(r"(select|exclude)[(][\\](\d+)", '\\1(dDA, m.start(\\2), m.group(\\2)', s)
    s = re.sub(r"define[(][\\](\d+)", 'define(dDA, m.start(\\1)', s)
    s = re.sub(r"(select|exclude)[(][\\](\d+)", '\\1(dTokenPos, m.start(\\2), m.group(\\2)', s)
    s = re.sub(r"define[(][\\](\d+)", 'define(dTokenPos, m.start(\\1)', s)
    s = re.sub(r"(morph|morphex|displayInfo)[(][\\](\d+)", '\\1((m.start(\\2), m.group(\\2))', s)
    s = re.sub(r"(morph|morphex|displayInfo)[(]", '\\1(dDA, ', s)
    s = re.sub(r"(morph|morphex|displayInfo)[(]", '\\1(dTokenPos, ', s)
    s = re.sub(r"(sugg\w+|switch\w+)\(@", '\\1(m.group(i[4])', s)
    s = re.sub(r"word\(\s*1\b", 'nextword1(s, m.end()', s)                                  # word(1)
    s = re.sub(r"word\(\s*-1\b", 'prevword1(s, m.start()', s)                               # word(-1)
    s = re.sub(r"word\(\s*(\d)", 'nextword(s, m.end(), \\1', s)                             # word(n)
    s = re.sub(r"word\(\s*-(\d)", 'prevword(s, m.start(), \\1', s)                          # word(-n)
    s = re.sub(r"before\(\s*", 'look(s[:m.start()], ', s)                                   # before(s)
    s = re.sub(r"after\(\s*", 'look(s[m.end():], ', s)                                      # after(s)
    s = re.sub(r"textarea\(\s*", 'look(s, ', s)                                             # textarea(s)
    s = re.sub(r"before_chk1\(\s*", 'look_chk1(dDA, s[:m.start()], 0, ', s)                 # before_chk1(s)
    s = re.sub(r"after_chk1\(\s*", 'look_chk1(dDA, s[m.end():], m.end(), ', s)              # after_chk1(s)
    s = re.sub(r"textarea_chk1\(\s*", 'look_chk1(dDA, s, 0, ', s)                           # textarea_chk1(s)
    s = re.sub(r"before_chk1\(\s*", 'look_chk1(dTokenPos, s[:m.start()], 0, ', s)           # before_chk1(s)
    s = re.sub(r"after_chk1\(\s*", 'look_chk1(dTokenPos, s[m.end():], m.end(), ', s)        # after_chk1(s)
    s = re.sub(r"textarea_chk1\(\s*", 'look_chk1(dTokenPos, s, 0, ', s)                     # textarea_chk1(s)
    s = re.sub(r"/0", 'sx[m.start():m.end()]', s)                                           # /0
    s = re.sub(r"before0\(\s*", 'look(sx[:m.start()], ', s)                                 # before0(s)
    s = re.sub(r"after0\(\s*", 'look(sx[m.end():], ', s)                                    # after0(s)
    s = re.sub(r"textarea0\(\s*", 'look(sx, ', s)                                           # textarea0(s)
    s = re.sub(r"before0_chk1\(\s*", 'look_chk1(dDA, sx[:m.start()], 0, ', s)               # before0_chk1(s)
    s = re.sub(r"after0_chk1\(\s*", 'look_chk1(dDA, sx[m.end():], m.end(), ', s)            # after0_chk1(s)
    s = re.sub(r"textarea0_chk1\(\s*", 'look_chk1(dDA, sx, 0, ', s)                         # textarea0_chk1(s)
    s = re.sub(r"isEndOfNG\(\s*\)", 'isEndOfNG(dDA, s[m.end():], m.end())', s)              # isEndOfNG(s)
    s = re.sub(r"isNextNotCOD\(\s*\)", 'isNextNotCOD(dDA, s[m.end():], m.end())', s)        # isNextNotCOD(s)
    s = re.sub(r"isNextVerb\(\s*\)", 'isNextVerb(dDA, s[m.end():], m.end())', s)            # isNextVerb(s)
    s = re.sub(r"before0_chk1\(\s*", 'look_chk1(dTokenPos, sx[:m.start()], 0, ', s)         # before0_chk1(s)
    s = re.sub(r"after0_chk1\(\s*", 'look_chk1(dTokenPos, sx[m.end():], m.end(), ', s)      # after0_chk1(s)
    s = re.sub(r"textarea0_chk1\(\s*", 'look_chk1(dTokenPos, sx, 0, ', s)                   # textarea0_chk1(s)
    s = re.sub(r"isEndOfNG\(\s*\)", 'isEndOfNG(dTokenPos, s[m.end():], m.end())', s)        # isEndOfNG(s)
    s = re.sub(r"isNextNotCOD\(\s*\)", 'isNextNotCOD(dTokenPos, s[m.end():], m.end())', s)  # isNextNotCOD(s)
    s = re.sub(r"isNextVerb\(\s*\)", 'isNextVerb(dTokenPos, s[m.end():], m.end())', s)      # isNextVerb(s)
    s = re.sub(r"\bspell *[(]", '_oSpellChecker.isValid(', s)
    s = re.sub(r"[\\](\d+)", 'm.group(\\1)', s)
    return s


def uppercase (s, sLang):
    "(flag i is not enough): converts regex to uppercase regex: 'foo' becomes '[Ff][Oo][Oo]', but 'Bar' becomes 'B[Aa][Rr]'."

            nState = 4
        elif nState == 4:
            nState = 0
    return sUp


def countGroupInRegex (sRegex):
    "returns the number of groups in <sRegex>"
    try:
        return re.compile(sRegex).groups
    except:
        traceback.print_exc()
        print(sRegex)
    return 0


def createRule (s, nIdLine, sLang, bParagraph, dOptPriority):
    "returns rule as list [option name, regex, bCaseInsensitive, identifier, list of actions]"
    global dJSREGEXES
    global nRULEWITHOUTNAME

    #### OPTIONS
    sLineId = str(nIdLine) + ("p" if bParagraph else "s")
    sRuleId = sLineId

    #### GRAPH CALL
    if s.startswith("@@@@"):
        if bParagraph:
            print("Error. Graph call can be made only after the first pass (sentence by sentence)")
            exit()
        return ["@@@@", s[4:], sLineId]

    #### OPTIONS
    sOption = False         # False or [a-z0-9]+ name
    nPriority = 4           # Default is 4, value must be between 0 and 9
    tGroups = None          # code for groups positioning (only useful for JavaScript)
    cCaseMode = 'i'         # i: case insensitive,  s: case sensitive,  u: uppercasing allowed
    cWordLimitLeft = '['    # [: word limit, <: no specific limit
    cWordLimitRight = ']'   # ]: word limit, >: no specific limit
    m = re.match("^__(?P<borders_and_case>[[<]\\w[]>])(?P<option>/[a-zA-Z0-9]+|)(?P<ruleid>\\(\\w+\\)|)(?P<priority>![0-9]|)__ *", s)

    #### REGEX TRIGGER
    i = s.find(" <<-")
    if i == -1:
        print("# Error: no condition at line " + sLineId)
        return None
    sRegex = s[:i].strip()
    s = s[i+4:]
    

    # JS groups positioning codes
    m = re.search("@@\\S+", sRegex)
    if m:
        tGroups = jsconv.groupsPositioningCodeToList(sRegex[m.start()+2:])
        sRegex = sRegex[:m.start()].strip()
    # JS regex
    m = re.search("<js>.+</js>i?", sRegex)

        sRegex = sRegex.replace("(?i)", "")
        sRegex = uppercase(sRegex, sLang)
    else:
        print("# Unknown case mode [" + cCaseMode + "] at line " + sLineId)

    ## check regex
    try:
        z = re.compile(sRegex)
        re.compile(sRegex)
    except:
        print("# Regex error at line ", nIdLine)
        print(sRegex)
        traceback.print_exc()
        return None
    ## groups in non grouping parenthesis
    for x in re.finditer("\(\?:[^)]*\([[\w -]", sRegex):
    for x in re.finditer(r"\(\?:[^)]*\([[\w -]", sRegex):
        print("# Warning: groups inside non grouping parenthesis in regex at line " + sLineId)

    #### PARSE ACTIONS
    lActions = []
    nAction = 1
    for sAction in s.split(" <<- "):
        t = createAction(sRuleId + "_" + str(nAction), sAction, nGroup)
        nAction += 1
        if t:
            lActions.append(t)
    if not lActions:
        return None

    return [sOption, sRegex, bCaseInsensitive, sLineId, sRuleId, nPriority, lActions, tGroups]


def checkReferenceNumbers (sText, sActionId, nToken):
    "check if token references in <sText> greater than <nToken> (debugging)"
    for x in re.finditer(r"\\(\d+)", sText):
        if int(x.group(1)) > nToken:
            print("# Error in token index at line " + sActionId + " ("+str(nToken)+" tokens only)")
            print(sText)


def checkIfThereIsCode (sText, sActionId):
    "check if there is code in <sText> (debugging)"
    if re.search("[.]\\w+[(]|sugg\\w+[(]|\\([0-9]|\\[[0-9]", sText):
        print("# Warning at line " + sActionId + ":  This message looks like code. Line should probably begin with =")
        print(sText)


def createAction (sIdAction, sAction, nGroup):
    "returns an action to perform as a tuple (condition, action type, action[, iGroup [, message, URL ]])"
    global lFUNCTIONS

    m = re.search(r"([-~=>])(\d*|)>>", sAction)
    if not m:
        print("# No action at line " + sIdAction)
        return None

    #### CONDITION
    sCondition = sAction[:m.start()].strip()
    if sCondition:
        sCondition = prepareFunction(sCondition)
        lFUNCTIONS.append(("c_"+sIdAction, sCondition))
        for x in re.finditer("[.](?:group|start|end)[(](\d+)[)]", sCondition):
        lFUNCTIONS.append(("_c_"+sIdAction, sCondition))
        checkReferenceNumbers(sCondition, sIdAction, nGroup)
            if int(x.group(1)) > nGroup:
                print("# Error in groups in condition at line " + sIdAction + " ("+str(nGroup)+" groups only)")
        if ".match" in sCondition:
            print("# Error. JS compatibility. Don't use .match() in condition, use .search()")
        sCondition = "c_"+sIdAction
        sCondition = "_c_"+sIdAction
    else:
        sCondition = None

    #### iGroup / positioning
    iGroup = int(m.group(2)) if m.group(2) else 0
    if iGroup > nGroup:
        print("# Selected group > group number in regex at line " + sIdAction)
    

    #### ACTION
    sAction = sAction[m.end():].strip()
    cAction = m.group(1)
    if cAction == "-":
        ## error
        iMsg = sAction.find(" # ")
        if iMsg == -1:
            sMsg = "# Error. Error message not found."
            sURL = ""
            print(sMsg + " Action id: " + sIdAction)
        else:
            sMsg = sAction[iMsg+3:].strip()
            sAction = sAction[:iMsg].strip()
            sURL = ""
            mURL = re.search("[|] *(https?://.*)", sMsg)
            if mURL:
                sURL = mURL.group(1).strip()
                sMsg = sMsg[:mURL.start(0)].strip()
            checkReferenceNumbers(sMsg, sIdAction, nGroup)
            if sMsg[0:1] == "=":
                sMsg = prepareFunction(sMsg[1:])
                lFUNCTIONS.append(("m_"+sIdAction, sMsg))
                lFUNCTIONS.append(("_m_"+sIdAction, sMsg))
                for x in re.finditer("group[(](\d+)[)]", sMsg):
                    if int(x.group(1)) > nGroup:
                        print("# Error in groups in message at line " + sIdAction + " ("+str(nGroup)+" groups only)")
                sMsg = "=m_"+sIdAction
                sMsg = "=_m_"+sIdAction
            else:
                for x in re.finditer(r"\\(\d+)", sMsg):
                    if int(x.group(1)) > nGroup:
                        print("# Error in groups in message at line " + sIdAction + " ("+str(nGroup)+" groups only)")
                checkIfThereIsCode(sMsg, sIdAction)
                if re.search("[.]\\w+[(]", sMsg):
                    print("# Error in message at line " + sIdAction + ":  This message looks like code. Line should begin with =")
            

    checkReferenceNumbers(sAction, sIdAction, nGroup)
    if sAction[0:1] == "=" or cAction == "=":
        if "define" in sAction and not re.search(r"define\(\\\d+ *, *\[.*\] *\)", sAction):
            print("# Error in action at line " + sIdAction + ": second argument for define must be a list of strings")
        sAction = prepareFunction(sAction)
        sAction = sAction.replace("m.group(i[4])", "m.group("+str(iGroup)+")")
        for x in re.finditer("group[(](\d+)[)]", sAction):
            if int(x.group(1)) > nGroup:
                print("# Error in groups in replacement at line " + sIdAction + " ("+str(nGroup)+" groups only)")
    else:
        for x in re.finditer(r"\\(\d+)", sAction):
            if int(x.group(1)) > nGroup:
                print("# Error in groups in replacement at line " + sIdAction + " ("+str(nGroup)+" groups only)")
        if re.search("[.]\\w+[(]|sugg\\w+[(]", sAction):
            print("# Error in action at line " + sIdAction + ":  This action looks like code. Line should begin with =")
        checkIfThereIsCode(sAction, sIdAction)

    if cAction == ">":
        ## no action, break loop if condition is False
        return [sCondition, cAction, ""]

    if not sAction:
        print("# Error in action at line " + sIdAction + ":  This action is empty.")
        return None

    if cAction == "-":
        ## error detected --> suggestion
        if not sAction:
            print("# Error in action at line " + sIdAction + ":  This action is empty.")
        if sAction[0:1] == "=":
            lFUNCTIONS.append(("s_"+sIdAction, sAction[1:]))
            sAction = "=s_"+sIdAction
            lFUNCTIONS.append(("_s_"+sIdAction, sAction[1:]))
            sAction = "=_s_"+sIdAction
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        if not sMsg:
            print("# Error in action at line " + sIdAction + ":  the message is empty.")
        return [sCondition, cAction, sAction, iGroup, sMsg, sURL]
    elif cAction == "~":
        ## text processor
        if not sAction:
            print("# Error in action at line " + sIdAction + ":  This action is empty.")
        if sAction[0:1] == "=":
            lFUNCTIONS.append(("p_"+sIdAction, sAction[1:]))
            sAction = "=p_"+sIdAction
            lFUNCTIONS.append(("_p_"+sIdAction, sAction[1:]))
            sAction = "=_p_"+sIdAction
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        return [sCondition, cAction, sAction, iGroup]
    elif cAction == "=":
        ## disambiguator
        if sAction[0:1] == "=":
            sAction = sAction[1:]
        if not sAction:
            print("# Error in action at line " + sIdAction + ":  This action is empty.")
        lFUNCTIONS.append(("d_"+sIdAction, sAction))
        sAction = "d_"+sIdAction
        if "define" in sAction and not re.search(r"define\(dTokenPos, *m\.start.*, \[.*\] *\)", sAction):
            print("# Error in action at line " + sIdAction + ": second argument for define must be a list of strings")
            print(sAction)
        lFUNCTIONS.append(("_d_"+sIdAction, sAction))
        sAction = "_d_"+sIdAction
        return [sCondition, cAction, sAction]
    elif cAction == ">":
        ## no action, break loop if condition is False
        return [sCondition, cAction, ""]
    else:
        print("# Unknown action at line " + sIdAction)
        return None


def _calcRulesStats (lRules):
    "count rules and actions"
    d = {'=':0, '~': 0, '-': 0, '>': 0}
    for aRule in lRules:
        if aRule[0] != "@@@@":
        for aAction in aRule[6]:
            d[aAction[1]] = d[aAction[1]] + 1
            for aAction in aRule[6]:
                d[aAction[1]] = d[aAction[1]] + 1
    return (d, len(lRules))


def displayStats (lParagraphRules, lSentenceRules):
    "display rules numbers"
    print("  {:>18} {:>18} {:>18} {:>18}".format("DISAMBIGUATOR", "TEXT PROCESSOR", "GRAMMAR CHECKING", "REGEX"))
    d, nRule = _calcRulesStats(lParagraphRules)
    print("§ {:>10} actions {:>10} actions {:>10} actions  in {:>8} rules".format(d['='], d['~'], d['-'], nRule))
    d, nRule = _calcRulesStats(lSentenceRules)
    print("s {:>10} actions {:>10} actions {:>10} actions  in {:>8} rules".format(d['='], d['~'], d['-'], nRule))

            m = re.match("OPTGROUP/([a-z0-9]+):(.+)$", sLine)
            lStructOpt.append( (m.group(1), list(map(str.split, m.group(2).split(",")))) )
        elif sLine.startswith("OPTSOFTWARE:"):
            lOpt = [ [s, {}]  for s in sLine[12:].strip().split() ]  # don’t use tuples (s, {}), because unknown to JS
        elif sLine.startswith("OPT/"):
            m = re.match("OPT/([a-z0-9]+):(.+)$", sLine)
            for i, sOpt in enumerate(m.group(2).split()):
                lOpt[i][1][m.group(1)] =  eval(sOpt)
                lOpt[i][1][m.group(1)] = eval(sOpt)
        elif sLine.startswith("OPTPRIORITY/"):
            m = re.match("OPTPRIORITY/([a-z0-9]+): *([0-9])$", sLine)
            dOptPriority[m.group(1)] = int(m.group(2))
        elif sLine.startswith("OPTLANG/"):
            m = re.match("OPTLANG/([a-z][a-z](?:_[A-Z][A-Z]|)):(.+)$", sLine)
            sLang = m.group(1)[:2]
            dOptLabel[sLang] = { "__optiontitle__": m.group(2).strip() }

    print("  options defined for: " + ", ".join([ t[0] for t in lOpt ]))
    dOptions = { "lStructOpt": lStructOpt, "dOptLabel": dOptLabel, "sDefaultUILang": sDefaultUILang }
    dOptions.update({ "dOpt"+k: v  for k, v in lOpt })
    return dOptions, dOptPriority


def printBookmark (nLevel, sComment, nLine):
    "print bookmark within the rules file"
    print("  {:>6}:  {}".format(nLine, "  " * nLevel + sComment))


def make (spLang, sLang, bJavaScript):
    "compile rules, returns a dictionary of values"
    # for clarity purpose, don’t create any file here

    print("> read rules file...")
    try:
        lRules = open(spLang + "/rules.grx", 'r', encoding="utf-8").readlines()
    except:
        print("Error. Rules file in project [" + sLang + "] not found.")
        exit()

    # removing comments, zeroing empty lines, creating definitions, storing tests, merging rule lines
    print("  parsing rules...")
    global dDEF
    lLine = []
    lRuleLine = []
    lTest = []
    lOpt = []
    zBookmark = re.compile("^!!+")
    zGraphLink = re.compile(r"^@@@@GRAPHLINK>(\w+)@@@@")
    bGraph = False
    lGraphRule = []

    for i, sLine in enumerate(lRules, 1):
        if sLine.startswith('#END'):
            # arbitrary end
            printBookmark(0, "BREAK BY #END", i)
            break
        elif sLine.startswith("#"):
            # comment
            pass
        elif sLine.startswith("@@@@"):
        elif sLine.startswith("DEF:"):
            m = re.match(r"^@@@@GRAPHLINK>(\w+)@@@@", sLine.strip())
            if m:
            # definition
                #lRuleLine.append(["@GRAPHLINK", m.group(1)])
                printBookmark(1, "@GRAPHLINK: " + m.group(1), i)
        elif sLine.startswith("DEF:"):
            m = re.match("DEF: +([a-zA-Z_][a-zA-Z_0-9]*) +(.+)$", sLine.strip())
            if m:
                dDEF["{"+m.group(1)+"}"] = m.group(2)
            else:
                print("Error in definition: ", end="")
                print(sLine.strip())
        elif sLine.startswith("TEST:"):
            # test
            lTest.append("{:<8}".format(i) + "  " + sLine[5:].strip())
        elif sLine.startswith("TODO:"):
            # todo
            pass
        elif sLine.startswith(("OPTGROUP/", "OPTSOFTWARE:", "OPT/", "OPTLANG/", "OPTDEFAULTUILANG:", "OPTLABEL/", "OPTPRIORITY/")):
            # options
            lOpt.append(sLine)
        elif re.match("[  \t]*$", sLine):
            pass
        elif sLine.startswith("!!"):
            m = zBookmark.search(sLine)
            # bookmark
            m = re.match("!!+", sLine)
            nExMk = len(m.group(0))
            if sLine[nExMk:].strip():
                printBookmark(nExMk-2, sLine[nExMk:].strip(), i)
        # Graph rules
        elif sLine.startswith("@@@@GRAPH:"):
            # rules graph call
            m = re.match(r"@@@@GRAPH: *(\w+)", sLine.strip())
            if m:
                printBookmark(1, "@GRAPH: " + m.group(1), i)
                lRuleLine.append([i, "@@@@"+m.group(1)])
                bGraph = True
            lGraphRule.append([i, sLine])
            bGraph = True
        elif sLine.startswith("@@@@END_GRAPH"):
            #lGraphRule.append([i, sLine])
            bGraph = False
        elif re.match("@@@@ *$", sLine):
            pass
        elif bGraph:
            lGraphRule.append([i, sLine])
        # Regex rules
        elif re.match("[  \t]*$", sLine):
            # empty line
            pass
        elif sLine.startswith(("    ", "\t")):
            # rule (continuation)
            lRuleLine[len(lRuleLine)-1][1] += " " + sLine.strip()
            lRuleLine[-1][1] += " " + sLine.strip()
        else:
            # new rule
            lRuleLine.append([i, sLine.strip()])

    # generating options files
    print("  parsing options...")
    try:
        dOptions, dOptPriority = prepareOptions(lOpt)
    except:

                        lSentenceRulesJS.append(jsconv.pyRuleToJS(aRule, dJSREGEXES, sWORDLIMITLEFT))

    # creating file with all functions callable by rules
    print("  creating callables...")
    sPyCallables = "# generated code, do not edit\n"
    sJSCallables = "// generated code, do not edit\nconst oEvalFunc = {\n"
    for sFuncName, sReturn in lFUNCTIONS:
        cType = sFuncName[0:1]
        if cType == "c": # condition
            sParams = "s, sx, m, dDA, sCountry, bCondMemo"
        elif cType == "m": # message
        if sFuncName.startswith("_c_"): # condition
            sParams = "s, sx, m, dTokenPos, sCountry, bCondMemo"
        elif sFuncName.startswith("_m_"): # message
            sParams = "s, m"
        elif cType == "s": # suggestion
        elif sFuncName.startswith("_s_"): # suggestion
            sParams = "s, m"
        elif cType == "p": # preprocessor
        elif sFuncName.startswith("_p_"): # preprocessor
            sParams = "s, m"
        elif cType == "d": # disambiguator
            sParams = "s, m, dDA"
        elif sFuncName.startswith("_d_"): # disambiguator
            sParams = "s, m, dTokenPos"
        else:
            print("# Unknown function type in [" + sFuncName + "]")
            continue
        sPyCallables += "def {} ({}):\n".format(sFuncName, sParams)
        sPyCallables += "    return " + sReturn + "\n"
        sJSCallables += "    {}: function ({})".format(sFuncName, sParams) + " {\n"
        sJSCallables += "        return " + jsconv.py2js(sReturn) + ";\n"
        sJSCallables += "    },\n"
    sJSCallables += "}\n"

    displayStats(lParagraphRules, lSentenceRules)

    print("Unnamed rules: " + str(nRULEWITHOUTNAME))

    d = { "callables": sPyCallables,
          "callablesJS": sJSCallables,
          "gctests": sGCTests,
          "gctestsJS": sGCTestsJS,
          "paragraph_rules": mergeRulesByOption(lParagraphRules),
          "sentence_rules": mergeRulesByOption(lSentenceRules),
          "paragraph_rules_JS": jsconv.writeRulesToJSArray(mergeRulesByOption(lParagraphRulesJS)),
          "sentence_rules_JS": jsconv.writeRulesToJSArray(mergeRulesByOption(lSentenceRulesJS)) }
    d.update(dOptions)
    dVars = {   "callables": sPyCallables,
                "callablesJS": sJSCallables,
                "gctests": sGCTests,
                "gctestsJS": sGCTestsJS,
                "paragraph_rules": mergeRulesByOption(lParagraphRules),
                "sentence_rules": mergeRulesByOption(lSentenceRules),
                "paragraph_rules_JS": jsconv.writeRulesToJSArray(mergeRulesByOption(lParagraphRulesJS)),
                "sentence_rules_JS": jsconv.writeRulesToJSArray(mergeRulesByOption(lSentenceRulesJS)) }
    dVars.update(dOptions)

    # compile graph rules
    dVars2 = crg.make(lGraphRule, dDEF, sLang, bJavaScript)
    dVars.update(dVars2)

    return d
    return dVars

"""
Grammalecte: compile rules
Create a Direct Acyclic Rule Graphs (DARGs)
"""

import re
import traceback
import json

import darg


dACTIONS = {}
dFUNCTIONS = {}


def prepareFunction (s, bTokenValue=False):
    "convert simple rule syntax to a string of Python code"
    s = s.replace("__also__", "bCondMemo")
    s = s.replace("__else__", "not bCondMemo")
    s = re.sub(r"(morph|analyse|displayInfo)[(]\\(\d+)", 'g_\\1(lToken[\\2+nTokenOffset]', s)
    s = re.sub(r"(select|exclude|define)[(][\\](\d+)", 'g_\\1(lToken[\\2+nTokenOffset], dTags', s)
    s = re.sub(r"(tag_before|tag_after)[(][\\](\d+)", 'g_\\1(lToken[\\2+nTokenOffset], dTags', s)
    s = re.sub(r"(switchGender|has(?:Mas|Fem)Form)[(]\\(\d+)", '\\1(lToken[\\2+nTokenOffset]["sValue"]', s)
    s = re.sub(r"(morph|analyse)\(>1", 'g_\\1(lToken[nLastToken+1]', s)                     # next token
    s = re.sub(r"(morph|analyse)\(<1", 'g_\\1(lToken[nTokenOffset]', s)                     # previous token
    s = re.sub(r"[\\](\d+)\.is(upper|lower|title)\(\)", 'lToken[\\1+nTokenOffset]["sValue"].is\\2()', s)
    s = re.sub(r"\bspell *[(]", '_oSpellChecker.isValid(', s)
    s = re.sub(r"\bbefore\(\s*", 'look(sSentence[:lToken[1+nTokenOffset]["nStart"]], ', s)          # before(s)
    s = re.sub(r"\bafter\(\s*", 'look(sSentence[lToken[nLastToken]["nEnd"]:], ', s)                 # after(s)
    s = re.sub(r"\bbefore0\(\s*", 'look(sSentence0[:lToken[1+nTokenOffset]["nStart"]], ', s)        # before0(s)
    s = re.sub(r"\bafter0\(\s*", 'look(sSentence[lToken[nLastToken]["nEnd"]:], ', s)                # after0(s)
    if bTokenValue:
        # token values are used as parameter
        s = re.sub(r"[\\](\d+)", 'lToken[\\1+nTokenOffset]["sValue"]', s)
    else:
        # tokens used as parameter
        s = re.sub(r"[\\](\d+)", 'lToken[\\1+nTokenOffset]', s)
    return s


def genTokenLines (sTokenLine, dDef):
    "tokenize a string and return a list of lines of tokens"
    lToken = sTokenLine.split()
    lTokenLines = None
    for sToken in lToken:
        # optional token?
        bNullPossible = sToken.startswith("?") and sToken.endswith("¿")
        if bNullPossible:
            sToken = sToken[1:-1]
        # token with definition?
        if sToken.startswith("({") and sToken.endswith("})") and sToken[1:-1] in dDef:
            sToken = "(" + dDef[sToken[1:-1]] + ")"
        elif sToken.startswith("{") and sToken.endswith("}") and sToken in dDef:
            sToken = dDef[sToken]
        if ( (sToken.startswith("[") and sToken.endswith("]")) or (sToken.startswith("([") and sToken.endswith("])")) ):
            # multiple token
            bSelectedGroup = sToken.startswith("(") and sToken.endswith(")")
            if bSelectedGroup:
                sToken = sToken[1:-1]
            lNewToken = sToken[1:-1].split("|")
            if not lTokenLines:
                lTokenLines = [ [s]  for s  in lNewToken ]
                if bNullPossible:
                    lTokenLines.extend([ []  for i  in range(len(lNewToken)+1) ])
            else:
                lNewTemp = []
                if bNullPossible:
                    for aRule in lTokenLines:
                        for sElem in lNewToken:
                            aNewRule = list(aRule)
                            aNewRule.append(sElem)
                            lNewTemp.append(aNewRule)
                else:
                    sElem1 = lNewToken.pop(0)
                    for aRule in lTokenLines:
                        for sElem in lNewToken:
                            aNewRule = list(aRule)
                            aNewRule.append("(" + sElem + ")"  if bSelectedGroup  else sElem)
                            lNewTemp.append(aNewRule)
                        aRule.append("(" + sElem1 + ")"  if bSelectedGroup  else sElem1)
                lTokenLines.extend(lNewTemp)
        else:
            # simple token
            if not lTokenLines:
                lTokenLines = [[sToken], []]  if bNullPossible  else [[sToken]]
            else:
                if bNullPossible:
                    lNewTemp = []
                    for aRule in lTokenLines:
                        lNew = list(aRule)
                        lNew.append(sToken)
                        lNewTemp.append(lNew)
                    lTokenLines.extend(lNewTemp)
                else:
                    for aRule in lTokenLines:
                        aRule.append(sToken)
    for aRule in lTokenLines:
        yield aRule


def createRule (iLine, sRuleName, sTokenLine, iActionBlock, sActions, nPriority, dDef):
    "generator: create rule as list"
    # print(iLine, "//", sRuleName, "//", sTokenLine, "//", sActions, "//", nPriority)
    for lToken in genTokenLines(sTokenLine, dDef):
        # Calculate positions
        dPos = {}   # key: iGroup, value: iToken
        iGroup = 0
        for i, sToken in enumerate(lToken):
            if sToken.startswith("(") and sToken.endswith(")"):
                lToken[i] = sToken[1:-1]
                iGroup += 1
                dPos[iGroup] = i + 1    # we add 1, for we count tokens from 1 to n (not from 0)

        # Parse actions
        for iAction, sAction in enumerate(sActions.split(" <<- "), 1):
            sAction = sAction.strip()
            if sAction:
                sActionId = sRuleName + "__b" + str(iActionBlock) + "_a" + str(iAction) + "_" + str(len(lToken))
                aAction = createAction(sActionId, sAction, nPriority, len(lToken), dPos)
                if aAction:
                    dACTIONS[sActionId] = aAction
                    lResult = list(lToken)
                    lResult.extend(["##"+str(iLine), sActionId])
                    yield lResult


def changeReferenceToken (sText, dPos):
    "change group reference in <sText> with values in <dPos>"
    for i in range(len(dPos), 0, -1):
        sText = sText.replace("\\"+str(i), "\\"+str(dPos[i]))
    return sText


def checkTokenNumbers (sText, sActionId, nToken):
    "check if token references in <sText> greater than <nToken> (debugging)"
    for x in re.finditer(r"\\(\d+)", sText):
        if int(x.group(1)) > nToken:
            print("# Error in token index at line " + sActionId + " ("+str(nToken)+" tokens only)")
            print(sText)


def checkIfThereIsCode (sText, sActionId):
    "check if there is code in <sText> (debugging)"
    if re.search("[.]\\w+[(]|sugg\\w+[(]|\\([0-9]|\\[[0-9]", sText):
        print("# Warning at line " + sActionId + ":  This message looks like code. Line should probably begin with =")
        print(sText)


def createAction (sActionId, sAction, nPriority, nToken, dPos):
    "create action rule as a list"
    # Option
    sOption = False
    m = re.match("/(\\w+)/", sAction)
    if m:
        sOption = m.group(1)
        sAction = sAction[m.end():].strip()
    # valid action?
    m = re.search("(?P<action>[-~=/])(?P<start>\\d+|)(?P<end>:\\d+|)>> ", sAction)
    if not m:
        print(" # Error. No action found at: ", sActionId)
        print("   ==", sAction, "==")
        return None
    # Condition
    sCondition = sAction[:m.start()].strip()
    if sCondition:
        sCondition = prepareFunction(sCondition)
        sCondition = changeReferenceToken(sCondition, dPos)
        dFUNCTIONS["_g_c_"+sActionId] = sCondition
        sCondition = "_g_c_"+sActionId
    else:
        sCondition = ""
    # Action
    cAction = m.group("action")
    sAction = sAction[m.end():].strip()
    sAction = changeReferenceToken(sAction, dPos)
    if not m.group("start"):
        iStartAction = 1
        iEndAction = 0
    else:
        iStartAction = int(m.group("start"))
        iEndAction = int(m.group("end")[1:])  if m.group("end")  else iStartAction
    if dPos:
        try:
            iStartAction = dPos[iStartAction]
            iEndAction = dPos[iEndAction]
        except:
            print("# Error. Wrong groups in: " + sActionId)

    if cAction == "-":
        ## error
        iMsg = sAction.find(" # ")
        if iMsg == -1:
            sMsg = "# Error. Error message not found."
            sURL = ""
            print(sMsg + " Action id: " + sActionId)
        else:
            sMsg = sAction[iMsg+3:].strip()
            sAction = sAction[:iMsg].strip()
            sURL = ""
            mURL = re.search("[|] *(https?://.*)", sMsg)
            if mURL:
                sURL = mURL.group(1).strip()
                sMsg = sMsg[:mURL.start(0)].strip()
            checkTokenNumbers(sMsg, sActionId, nToken)
            if sMsg[0:1] == "=":
                sMsg = prepareFunction(sMsg[1:], True)
                dFUNCTIONS["g_m_"+sActionId] = sMsg
                sMsg = "=g_m_"+sActionId
            else:
                checkIfThereIsCode(sMsg, sActionId)

    # checking consistancy
    checkTokenNumbers(sAction, sActionId, nToken)

    if cAction == ">":
        ## no action, break loop if condition is False
        return [sOption, sCondition, cAction, ""]

    if not sAction:
        print("# Error in action at line " + sActionId + ":  This action is empty.")

    if sAction[0:1] != "=":
        checkIfThereIsCode(sAction, sActionId)

    if cAction == "-":
        ## error detected --> suggestion
        if sAction[0:1] == "=":
            sAction = prepareFunction(sAction, True)
            dFUNCTIONS["_g_s_"+sActionId] = sAction[1:]
            sAction = "=_g_s_"+sActionId
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        if not sMsg:
            print("# Error in action at line " + sActionId + ":  The message is empty.")
        return [sOption, sCondition, cAction, sAction, iStartAction, iEndAction, nPriority, sMsg, sURL]
    elif cAction == "~":
        ## text processor
        if sAction[0:1] == "=":
            dFUNCTIONS["_g_p_"+sActionId] = sAction[1:]
            sAction = "=_g_p_"+sActionId
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        return [sOption, sCondition, cAction, sAction, iStartAction, iEndAction]
    elif cAction == "/":
        ## tags
        return [sOption, sCondition, cAction, sAction, iStartAction, iEndAction]
    elif cAction == "=":
        ## disambiguator
        if sAction[0:1] == "=":
            sAction = sAction[1:]
        if "define" in sAction and not re.search(r"define\(\\\d+ *, *\[.*\] *\)", sAction):
            print("# Error in action at line " + sActionId + ": second argument for <define> must be a list of strings")
        sAction = prepareFunction(sAction)
        dFUNCTIONS["_g_d_"+sActionId] = sAction
        sAction = "_g_d_"+sActionId
        return [sOption, sCondition, cAction, sAction]
    else:
        print("# Unknown action at line " + sActionId)
        return None


def make (lRule, dDef, sLang, bJavaScript):
    "compile rules, returns a dictionary of values"
    # for clarity purpose, don’t create any file here

    # removing comments, zeroing empty lines, creating definitions, storing tests, merging rule lines
    print("  parsing rules...")
    lTokenLine = []
    sActions = ""
    nPriority = 4
    dAllGraph = {}
    sGraphName = ""
    iActionBlock = 0

    for i, sLine in lRule:
        sLine = sLine.rstrip()
        if "\t" in sLine:
            # tabulation not allowed
            print("Error. Tabulation at line: ", i)
            exit()
        elif sLine.startswith("@@@@GRAPH: "):
            # rules graph call
            m = re.match(r"@@@@GRAPH: *(\w+)", sLine.strip())
            if m:
                sGraphName = m.group(1)
                if sGraphName in dAllGraph:
                    print("Error. Group name " + sGraphName + " already exists.")
                    exit()
                dAllGraph[sGraphName] = []
            else:
                print("Error. Graph name not found at line", i)
                exit()
        elif sLine.startswith("__") and sLine.endswith("__"):
            # new rule group
            m = re.match("__(\\w+)(!\\d|)__", sLine)
            if m:
                sRuleName = m.group(1)
                iActionBlock = 1
                nPriority = int(m.group(2)[1:]) if m.group(2)  else 4
            else:
                print("Error at rule group: ", sLine, " -- line:", i)
                break
        elif re.search("^    +<<- ", sLine) or sLine.startswith("        ") \
                or re.search("^    +#", sLine) or re.search(r"^    [-~=>/](?:\d(?::\d+|)|)>> ", sLine) :
            # actions
            sActions += " " + sLine.strip()
        elif re.match("[  ]*$", sLine):
            # empty line to end merging
            if not lTokenLine:
                continue
            if not sActions:
                print("Error. No action found at line:", i)
                exit()
            if not sGraphName:
                print("Error. All rules must belong to a named graph. Line: ", i)
                exit()
            for j, sTokenLine in lTokenLine:
                dAllGraph[sGraphName].append((j, sRuleName, sTokenLine, iActionBlock, sActions, nPriority))
            lTokenLine.clear()
            sActions = ""
            iActionBlock += 1
        elif sLine.startswith(("    ")):
            # tokens
            lTokenLine.append([i, sLine.strip()])
        else:
            print("Unknown line:")
            print(sLine)

    # processing rules
    print("  preparing rules...")
    for sGraphName, lRuleLine in dAllGraph.items():
        lPreparedRule = []
        for i, sRuleGroup, sTokenLine, iActionBlock, sActions, nPriority in lRuleLine:
            for lRule in createRule(i, sRuleGroup, sTokenLine, iActionBlock, sActions, nPriority, dDef):
                lPreparedRule.append(lRule)
        # Graph creation
        oDARG = darg.DARG(lPreparedRule, sLang)
        dAllGraph[sGraphName] = oDARG.createGraph()
        # Debugging
        #print("\nGRAPH:", sGraphName)
        #for e in lPreparedRule:
        #    print(e)
        #for k, v in dAllGraph[sGraphName].items():
        #    print(k, "\t", v)

    # creating file with all functions callable by rules
    print("  creating callables...")
    sPyCallables = "# generated code, do not edit\n"
    #sJSCallables = "// generated code, do not edit\nconst oEvalFunc = {\n"
    for sFuncName, sReturn in dFUNCTIONS.items():
        if sFuncName.startswith("_g_c_"): # condition
            sParams = "lToken, nTokenOffset, nLastToken, sCountry, bCondMemo, dTags, sSentence, sSentence0"
        elif sFuncName.startswith("g_m_"): # message
            sParams = "lToken, nTokenOffset"
        elif sFuncName.startswith("_g_s_"): # suggestion
            sParams = "lToken, nTokenOffset"
        elif sFuncName.startswith("_g_p_"): # preprocessor
            sParams = "lToken"
        elif sFuncName.startswith("_g_d_"): # disambiguator
            sParams = "lToken, nTokenOffset"
        else:
            print("# Unknown function type in [" + sFuncName + "]")
            continue
        sPyCallables += "def {} ({}):\n".format(sFuncName, sParams)
        sPyCallables += "    return " + sReturn + "\n"
        #sJSCallables += "    {}: function ({})".format(sFuncName, sParams) + " {\n"
        #sJSCallables += "        return " + jsconv.py2js(sReturn) + ";\n"
        #sJSCallables += "    },\n"
    #sJSCallables += "}\n"

    # Debugging
    if False:
        print("\nActions:")
        for sActionName, aAction in dACTIONS.items():
            print(sActionName, aAction)
        print("\nFunctions:")
        print(sPyCallables)

    # Result
    return {
        "graph_callables": sPyCallables,
        "rules_graphs": dAllGraph,
        "rules_actions": dACTIONS
    }

"""
# Convert Python code to JavaScript code
Convert Python code and regexes to JavaScript code
"""

import copy
import re
import json


def py2js (sCode):

        sRegex = sRegex + "i"
    if not lNegLookBeforeRegex:
        lNegLookBeforeRegex = None
    return (sRegex, lNegLookBeforeRegex)


def pyRuleToJS (lRule, dJSREGEXES, sWORDLIMITLEFT):
    "modify Python rules -> JS rules"
    lRuleJS = copy.deepcopy(lRule)
    # graph rules
    if lRuleJS[0] == "@@@@":
        return lRuleJS
    del lRule[-1] # tGroups positioning codes are useless for Python
    # error messages
    for aAction in lRuleJS[6]:
        if aAction[1] == "-":
            aAction[2] = aAction[2].replace(" ", " ") # nbsp --> nnbsp
            aAction[4] = aAction[4].replace("« ", "« ").replace(" »", " »").replace(" :", " :").replace(" :", " :")
    # js regexes
    lRuleJS[1], lNegLookBehindRegex = regex2js(dJSREGEXES.get(lRuleJS[3], lRuleJS[1]), sWORDLIMITLEFT)
    lRuleJS.append(lNegLookBehindRegex)
    return lRuleJS


def writeRulesToJSArray (lRules):
    "create rules as a string of arrays (to be bundled in a JSON string)"
    sArray = "[\n"
    for sOption, aRuleGroup in lRules:
        if sOption != "@@@@":
        sArray += '  ["' + sOption + '", [\n'  if sOption  else  "  [false, [\n"
        for sRegex, bCaseInsensitive, sLineId, sRuleId, nPriority, lActions, aGroups, aNegLookBehindRegex in aRuleGroup:
            sArray += '    [' + sRegex + ", "
            sArray += "true, " if bCaseInsensitive  else "false, "
            sArray += '"' + sLineId + '", '
            sArray += '"' + sRuleId + '", '
            sArray += str(nPriority) + ", "
            sArray += json.dumps(lActions, ensure_ascii=False) + ", "
            sArray += json.dumps(aGroups, ensure_ascii=False) + ", "
            sArray += json.dumps(aNegLookBehindRegex, ensure_ascii=False) + "],\n"
        sArray += "  ]],\n"
            sArray += '  ["' + sOption + '", [\n'  if sOption  else  "  [false, [\n"
            for sRegex, bCaseInsensitive, sLineId, sRuleId, nPriority, lActions, aGroups, aNegLookBehindRegex in aRuleGroup:
                sArray += '    [' + sRegex + ", "
                sArray += "true, " if bCaseInsensitive  else "false, "
                sArray += '"' + sLineId + '", '
                sArray += '"' + sRuleId + '", '
                sArray += str(nPriority) + ", "
                sArray += json.dumps(lActions, ensure_ascii=False) + ", "
                sArray += json.dumps(aGroups, ensure_ascii=False) + ", "
                sArray += json.dumps(aNegLookBehindRegex, ensure_ascii=False) + "],\n"
            sArray += "  ]],\n"
        else:
            sArray += '  ["' + sOption + '", [\n'
            for sGraphName, sLineId in aRuleGroup:
                sArray += '    ["' + sGraphName + '", "' + sLineId + '"],\n"'
            sArray += "  ]],\n"
    sArray += "]"
    return sArray


def groupsPositioningCodeToList (sGroupsPositioningCode):
    "convert <sGroupsPositioningCode> to a list of codes (numbers or strings)"
    if not sGroupsPositioningCode:
        return None
    return [ int(sCode)  if sCode.isdigit() or (sCode[0:1] == "-" and sCode[1:].isdigit())  else sCode \
             for sCode in sGroupsPositioningCode.split(",") ]

#!python3

"""
RULE GRAPH BUILDER
"""

# by Olivier R.
# License: MPL 2


from graphspell.progressbar import ProgressBar


class DARG:
    """DIRECT ACYCLIC RULE GRAPH"""
    # This code is inspired from Steve Hanov’s DAWG, 2011. (http://stevehanov.ca/blog/index.php?id=115)

    def __init__ (self, lRule, sLangCode):
        print("===== Direct Acyclic Rule Graph - Minimal Acyclic Finite State Automaton =====")

        # Preparing DARG
        print(" > Preparing list of tokens")
        self.sLangCode = sLangCode
        self.nRule = len(lRule)
        self.aPreviousRule = []
        Node.resetNextId()
        self.oRoot = Node()
        self.lUncheckedNodes = []  # list of nodes that have not been checked for duplication.
        self.lMinimizedNodes = {}  # list of unique nodes that have been checked for duplication.
        self.nNode = 0
        self.nArc = 0

        # build
        lRule.sort()
        oProgBar = ProgressBar(0, len(lRule))
        for aRule in lRule:
            self.insert(aRule)
            oProgBar.increment(1)
        oProgBar.done()
        self.finish()
        self.countNodes()
        self.countArcs()
        self.displayInfo()

    # BUILD DARG
    def insert (self, aRule):
        "insert a new rule (tokens must be inserted in order)"
        if aRule < self.aPreviousRule:
            exit("# Error: tokens must be inserted in order.")

        # find common prefix between word and previous word
        nCommonPrefix = 0
        for i in range(min(len(aRule), len(self.aPreviousRule))):
            if aRule[i] != self.aPreviousRule[i]:
                break
            nCommonPrefix += 1

        # Check the lUncheckedNodes for redundant nodes, proceeding from last
        # one down to the common prefix size. Then truncate the list at that point.
        self._minimize(nCommonPrefix)

        # add the suffix, starting from the correct node mid-way through the graph
        if len(self.lUncheckedNodes) == 0:
            oNode = self.oRoot
        else:
            oNode = self.lUncheckedNodes[-1][2]

        iToken = nCommonPrefix
        for sToken in aRule[nCommonPrefix:]:
            oNextNode = Node()
            oNode.dArcs[sToken] = oNextNode
            self.lUncheckedNodes.append((oNode, sToken, oNextNode))
            if iToken == (len(aRule) - 2):
                oNode.bFinal = True
            iToken += 1
            oNode = oNextNode
        oNode.bFinal = True
        self.aPreviousRule = aRule

    def finish (self):
        "minimize unchecked nodes"
        self._minimize(0)

    def _minimize (self, downTo):
        # proceed from the leaf up to a certain point
        for i in range( len(self.lUncheckedNodes)-1, downTo-1, -1 ):
            oNode, sToken, oChildNode = self.lUncheckedNodes[i]
            if oChildNode in self.lMinimizedNodes:
                # replace the child with the previously encountered one
                oNode.dArcs[sToken] = self.lMinimizedNodes[oChildNode]
            else:
                # add the state to the minimized nodes.
                self.lMinimizedNodes[oChildNode] = oChildNode
            self.lUncheckedNodes.pop()

    def countNodes (self):
        "count nodes within the whole graph"
        self.nNode = len(self.lMinimizedNodes)

    def countArcs (self):
        "count arcs within the whole graph"
        self.nArc = 0
        for oNode in self.lMinimizedNodes:
            self.nArc += len(oNode.dArcs)

    def displayInfo (self):
        "display informations about the rule graph"
        print(" * {:<12} {:>16,}".format("Rules:", self.nRule))
        print(" * {:<12} {:>16,}".format("Nodes:", self.nNode))
        print(" * {:<12} {:>16,}".format("Arcs:", self.nArc))

    def createGraph (self):
        "create the graph as a dictionary"
        dGraph = { 0: self.oRoot.getNodeAsDict() }
        for oNode in self.lMinimizedNodes:
            sHashId = oNode.__hash__()
            if sHashId not in dGraph:
                dGraph[sHashId] = oNode.getNodeAsDict()
            else:
                print("Error. Double node… same id: ", sHashId)
                print(str(oNode.getNodeAsDict()))
        return dGraph



class Node:
    """Node of the rule graph"""

    NextId = 0

    def __init__ (self):
        self.i = Node.NextId
        Node.NextId += 1
        self.bFinal = False
        self.dArcs = {}          # key: arc value; value: a node

    @classmethod
    def resetNextId (cls):
        "reset to 0 the node counter"
        cls.NextId = 0

    def __str__ (self):
        # Caution! this function is used for hashing and comparison!
        cFinal = "1"  if self.bFinal  else "0"
        l = [cFinal]
        for (key, oNode) in self.dArcs.items():
            l.append(str(key))
            l.append(str(oNode.i))
        return "_".join(l)

    def __hash__ (self):
        # Used as a key in a python dictionary.
        return self.__str__().__hash__()

    def __eq__ (self, other):
        # Used as a key in a python dictionary.
        # Nodes are equivalent if they have identical arcs, and each identical arc leads to identical states.
        return self.__str__() == other.__str__()

    def getNodeAsDict (self):
        "returns the node as a dictionary structure"
        dNode = {}
        dReValue = {}
        dReMorph = {}
        dRule = {}
        dLemma = {}
        dMeta = {}
        for sArc, oNode in self.dArcs.items():
            if sArc.startswith("@") and len(sArc) > 1:
                dReMorph[sArc[1:]] = oNode.__hash__()
            elif sArc.startswith("~") and len(sArc) > 1:
                dReValue[sArc[1:]] = oNode.__hash__()
            elif sArc.startswith(">") and len(sArc) > 1:
                dLemma[sArc[1:]] = oNode.__hash__()
            elif sArc.startswith("*") and len(sArc) > 1:
                dMeta[sArc[1:]] = oNode.__hash__()
            elif sArc.startswith("##"):
                dRule[sArc[1:]] = oNode.__hash__()
            else:
                dNode[sArc] = oNode.__hash__()
        if dReValue:
            dNode["<re_value>"] = dReValue
        if dReMorph:
            dNode["<re_morph>"] = dReMorph
        if dLemma:
            dNode["<lemmas>"] = dLemma
        if dMeta:
            dNode["<meta>"] = dMeta
        if dRule:
            dNode["<rules>"] = dRule
        #if self.bFinal:
        #    dNode["<final>"] = 1
        return dNode

// data
let _sAppContext = "";                                  // what software is running
let _dOptions = null;
let _aIgnoredRules = new Set();
let _oSpellChecker = null;
let _dAnalyses = new Map();                             // cache for data from dictionary


var gc_engine = {

    //// Informations

    lang: "${lang}",

                var spellchecker = require("resource://grammalecte/graphspell/spellchecker.js");
                _oSpellChecker = new spellchecker.SpellChecker("${lang}", "", "${dic_main_filename_js}", "${dic_extended_filename_js}", "${dic_community_filename_js}", "${dic_personal_filename_js}");
            } else {
                _oSpellChecker = new SpellChecker("${lang}", sPath, "${dic_main_filename_js}", "${dic_extended_filename_js}", "${dic_community_filename_js}", "${dic_personal_filename_js}");
            }
            _sAppContext = sContext;
            _dOptions = gc_options.getOptions(sContext).gl_shallowCopy();     // duplication necessary, to be able to reset to default
            _oSpellChecker.activateStorage();
        }
        catch (e) {
            helpers.logerror(e);
        }
    },

    getSpellChecker: function () {

function displayInfo (dDA, aWord) {
    // for debugging: info of word
    if (!aWord) {
        helpers.echo("> nothing to find");
        return true;
    }
    if (!_dAnalyses.has(aWord[1]) && !_storeMorphFromFSA(aWord[1])) {
        helpers.echo("> not in FSA");
    let lMorph = _oSpellChecker.getMorph(aWord[1]);
    if (lMorph.length === 0) {
        helpers.echo("> not in dictionary");
        return true;
    }
    if (dDA.has(aWord[0])) {
        helpers.echo("DA: " + dDA.get(aWord[0]));
    }
    helpers.echo("FSA: " + _dAnalyses.get(aWord[1]));
    helpers.echo("FSA: " + lMorph);
    return true;
}

function _storeMorphFromFSA (sWord) {
    // retrieves morphologies list from _oSpellChecker -> _dAnalyses
    //helpers.echo("register: "+sWord + " " + _oSpellChecker.getMorph(sWord).toString())
    _dAnalyses.set(sWord, _oSpellChecker.getMorph(sWord));
    return !!_dAnalyses.get(sWord);
}

function morph (dDA, aWord, sPattern, bStrict=true, bNoWord=false) {
    // analyse a tuple (position, word), return true if sPattern in morphologies (disambiguation on)
    if (!aWord) {
        //helpers.echo("morph: noword, returns " + bNoWord);
        return bNoWord;
    }
    //helpers.echo("aWord: "+aWord.toString());
    if (!_dAnalyses.has(aWord[1]) && !_storeMorphFromFSA(aWord[1])) {
        return false;
    }
    let lMorph = dDA.has(aWord[0]) ? dDA.get(aWord[0]) : _dAnalyses.get(aWord[1]);
    let lMorph = dDA.has(aWord[0]) ? dDA.get(aWord[0]) : _oSpellChecker.getMorph(aWord[1]);
    //helpers.echo("lMorph: "+lMorph.toString());
    if (lMorph.length === 0) {
        return false;
    }
    //helpers.echo("***");
    if (bStrict) {
        return lMorph.every(s  =>  (s.search(sPattern) !== -1));
    }
    return lMorph.some(s  =>  (s.search(sPattern) !== -1));
}

function morphex (dDA, aWord, sPattern, sNegPattern, bNoWord=false) {
    // analyse a tuple (position, word), returns true if not sNegPattern in word morphologies and sPattern in word morphologies (disambiguation on)
    if (!aWord) {
        //helpers.echo("morph: noword, returns " + bNoWord);
        return bNoWord;
    }
    //helpers.echo("aWord: "+aWord.toString());
    if (!_dAnalyses.has(aWord[1]) && !_storeMorphFromFSA(aWord[1])) {
        return false;
    }
    let lMorph = dDA.has(aWord[0]) ? dDA.get(aWord[0]) : _dAnalyses.get(aWord[1]);
    let lMorph = dDA.has(aWord[0]) ? dDA.get(aWord[0]) : _oSpellChecker.getMorph(aWord[1]);
    //helpers.echo("lMorph: "+lMorph.toString());
    if (lMorph.length === 0) {
        return false;
    }
    //helpers.echo("***");
    // check negative condition
    if (lMorph.some(s  =>  (s.search(sNegPattern) !== -1))) {
        return false;
    }
    // search sPattern
    return lMorph.some(s  =>  (s.search(sPattern) !== -1));
}

function analyse (sWord, sPattern, bStrict=true) {
    // analyse a word, return true if sPattern in morphologies (disambiguation off)
    if (!_dAnalyses.has(sWord) && !_storeMorphFromFSA(sWord)) {
    let lMorph = _oSpellChecker.getMorph(sWord);
    if (lMorph.length === 0) {
        return false;
    }
    if (bStrict) {
        return _dAnalyses.get(sWord).every(s  =>  (s.search(sPattern) !== -1));
        return lMorph.every(s  =>  (s.search(sPattern) !== -1));
    }
    return _dAnalyses.get(sWord).some(s  =>  (s.search(sPattern) !== -1));
    return lMorph.some(s  =>  (s.search(sPattern) !== -1));
}

function analysex (sWord, sPattern, sNegPattern) {
    // analyse a word, returns True if not sNegPattern in word morphologies and sPattern in word morphologies (disambiguation off)
    if (!_dAnalyses.has(sWord) && !_storeMorphFromFSA(sWord)) {
    let lMorph = _oSpellChecker.getMorph(sWord);
    if (lMorph.length === 0) {
        return false;
    }
    // check negative condition
    if (_dAnalyses.get(sWord).some(s  =>  (s.search(sNegPattern) !== -1))) {
    if (lMorph.some(s  =>  (s.search(sNegPattern) !== -1))) {
        return false;
    }
    // search sPattern
    return _dAnalyses.get(sWord).some(s  =>  (s.search(sPattern) !== -1));
    return lMorph.some(s  =>  (s.search(sPattern) !== -1));
}

function stem (sWord) {
    // returns a list of sWord's stems
    if (!sWord) {
        return [];
    }
    if (!_dAnalyses.has(sWord) && !_storeMorphFromFSA(sWord)) {
        return [];
    }
    return _dAnalyses.get(sWord).map( s => s.slice(1, s.indexOf(" ")) );
}


//// functions to get text outside pattern scope

// warning: check compile_rules.py to understand how it works

function select (dDA, nPos, sWord, sPattern, lDefault=null) {
    if (!sWord) {
        return true;
    }
    if (dDA.has(nPos)) {
        return true;
    }
    if (!_dAnalyses.has(sWord) && !_storeMorphFromFSA(sWord)) {
    let lMorph = _oSpellChecker.getMorph(sWord);
        return true;
    }
    if (_dAnalyses.get(sWord).length === 1) {
    if (lMorph.length === 0  ||  lMorph.length === 1) {
        return true;
    }
    let lSelect = _dAnalyses.get(sWord).filter( sMorph => sMorph.search(sPattern) !== -1 );
    let lSelect = lMorph.filter( sMorph => sMorph.search(sPattern) !== -1 );
    if (lSelect.length > 0) {
        if (lSelect.length != _dAnalyses.get(sWord).length) {
        if (lSelect.length != lMorph.length) {
            dDA.set(nPos, lSelect);
        }
    } else if (lDefault) {
        dDA.set(nPos, lDefaul);
    }
    return true;
}

function exclude (dDA, nPos, sWord, sPattern, lDefault=null) {
    if (!sWord) {
        return true;
    }
    if (dDA.has(nPos)) {
        return true;
    }
    if (!_dAnalyses.has(sWord) && !_storeMorphFromFSA(sWord)) {
    let lMorph = _oSpellChecker.getMorph(sWord);
        return true;
    }
    if (_dAnalyses.get(sWord).length === 1) {
    if (lMorph.length === 0  ||  lMorph.length === 1) {
        return true;
    }
    let lSelect = _dAnalyses.get(sWord).filter( sMorph => sMorph.search(sPattern) === -1 );
    let lSelect = lMorph.filter( sMorph => sMorph.search(sPattern) === -1 );
    if (lSelect.length > 0) {
        if (lSelect.length != _dAnalyses.get(sWord).length) {
        if (lSelect.length != lMorph.length) {
            dDA.set(nPos, lSelect);
        }
    } else if (lDefault) {
        dDA.set(nPos, lDefault);
    }
    return true;
}

"""
Grammar checker
"""

from .grammar_checker import *

"""
# Grammalecte
Grammalecte, grammar checker
# Main class: wrapper
"""

import importlib
import json

from . import text


class GrammarChecker:
    "GrammarChecker: Wrapper for the grammar checker engine"

    def __init__ (self, sLangCode, sContext="Python"):
        self.sLangCode = sLangCode
        # Grammar checker engine
        self.gce = importlib.import_module("."+sLangCode, "grammalecte")
        self.gce.load(sContext)
        # Spell checker
        self.oSpellChecker = self.gce.getSpellChecker()
        # Lexicographer
        self.oLexicographer = None
        # Text formatter
        self.oTextFormatter = None

    def getGCEngine (self):
        "return the grammar checker object"
        return self.gce

    def getSpellChecker (self):
        "return the spell checker object"
        return self.oSpellChecker

    def getTextFormatter (self):
        "load and return the text formatter"
        if self.oTextFormatter == None:
            self.tf = importlib.import_module("."+self.sLangCode+".textformatter", "grammalecte")
        self.oTextFormatter = self.tf.TextFormatter()
        if self.oTextFormatter is None:
            tf = importlib.import_module("."+self.sLangCode+".textformatter", "grammalecte")
            self.oTextFormatter = tf.TextFormatter()
        return self.oTextFormatter

    def getLexicographer (self):
        "load and return the lexicographer"
        if self.oLexicographer == None:
            self.lxg = importlib.import_module("."+self.sLangCode+".lexicographe", "grammalecte")
        self.oLexicographer = self.lxg.Lexicographe(self.oSpellChecker)
        if self.oLexicographer is None:
            lxg = importlib.import_module("."+self.sLangCode+".lexicographe", "grammalecte")
            self.oLexicographer = lxg.Lexicographe(self.oSpellChecker)
        return self.oLexicographer

    def displayGCOptions (self):
        "display the grammar checker options"
        self.gce.displayOptions()

    def getParagraphErrors (self, sText, dOptions=None, bContext=False, bSpellSugg=False, bDebug=False):
        "returns a tuple: (grammar errors, spelling errors)"
        aGrammErrs = self.gce.parse(sText, "FR", bDebug=bDebug, dOptions=dOptions, bContext=bContext)
        aSpellErrs = self.oSpellChecker.parseParagraph(sText, bSpellSugg)
        return aGrammErrs, aSpellErrs

    def generateText (self, sText, bEmptyIfNoErrors=False, bSpellSugg=False, nWidth=100, bDebug=False):
        "[todo]"
        pass

    def generateTextAsJSON (self, sText, bContext=False, bEmptyIfNoErrors=False, bSpellSugg=False, bReturnText=False, bDebug=False):
        "[todo]"
        pass

    def generateParagraph (self, sText, dOptions=None, bEmptyIfNoErrors=False, bSpellSugg=False, nWidth=100, bDebug=False):
        "parse text and return a readable text with underline errors"
        aGrammErrs, aSpellErrs = self.getParagraphErrors(sText, dOptions, False, bSpellSugg, bDebug)
        if bEmptyIfNoErrors and not aGrammErrs and not aSpellErrs:
            return ""
        return text.generateParagraph(sText, aGrammErrs, aSpellErrs, nWidth)

    def generateParagraphAsJSON (self, iIndex, sText, dOptions=None, bContext=False, bEmptyIfNoErrors=False, bSpellSugg=False, bReturnText=False, lLineSet=None, bDebug=False):
        "parse text and return errors as a JSON string"
        aGrammErrs, aSpellErrs = self.getParagraphErrors(sText, dOptions, bContext, bSpellSugg, bDebug)
        aGrammErrs = list(aGrammErrs)
        if bEmptyIfNoErrors and not aGrammErrs and not aSpellErrs:
            return ""
        if lLineSet:
            aGrammErrs, aSpellErrs = text.convertToXY(aGrammErrs, aSpellErrs, lLineSet)
            return json.dumps({ "lGrammarErrors": aGrammErrs, "lSpellingErrors": aSpellErrs }, ensure_ascii=False)
        if bReturnText:
            return json.dumps({ "iParagraph": iIndex, "sText": sText, "lGrammarErrors": aGrammErrs, "lSpellingErrors": aSpellErrs }, ensure_ascii=False)
        return json.dumps({ "iParagraph": iIndex, "lGrammarErrors": aGrammErrs, "lSpellingErrors": aSpellErrs }, ensure_ascii=False)

"""
# Grammalecte
# Grammar checker engine
Grammalecte
Grammar checker engine
"""

import re
import sys
import os
import traceback
#import unicodedata
from itertools import chain

from ..graphspell.spellchecker import SpellChecker
from ..graphspell.echo import echo
from . import gc_options

from ..graphspell.tokenizer import Tokenizer
from .gc_rules_graph import dAllGraph, dRule

try:
    # LibreOffice / OpenOffice
    from com.sun.star.linguistic2 import SingleProofreadingError
    from com.sun.star.text.TextMarkupType import PROOFREADING
    from com.sun.star.beans import PropertyValue
    #import lightproof_handler_${implname} as opt
    _bWriterError = True
except ImportError:
    _bWriterError = False


__all__ = [ "lang", "locales", "pkg", "name", "version", "author", \
            "load", "parse", "getSpellChecker", \
            "setOption", "setOptions", "getOptions", "getDefaultOptions", "getOptionsLabels", "resetOptions", "displayOptions", \
            "ignoreRule", "resetIgnoreRules", "reactivateRule", "listRules", "displayRules" ]

author = "${author}"

_rules = None                               # module gc_rules

# data
_sAppContext = ""                           # what software is running
_dOptions = None
_oSpellChecker = None
_oTokenizer = None
_aIgnoredRules = set()

# functions
_createRegexError = None


#### Initialization

def load (sContext="Python"):
    "initialization of the grammar checker"
    global _oSpellChecker
    global _sAppContext
    global _dOptions
    global _oTokenizer
    global _createRegexError
    try:
        _oSpellChecker = SpellChecker("${lang}", "${dic_main_filename_py}", "${dic_extended_filename_py}", "${dic_community_filename_py}", "${dic_personal_filename_py}")
        _sAppContext = sContext
        _dOptions = dict(gc_options.getOptions(sContext))   # duplication necessary, to be able to reset to default
        _oTokenizer = _oSpellChecker.getTokenizer()
_oSpellChecker = None
_dAnalyses = {}                             # cache for data from dictionary
        _oSpellChecker.activateStorage()
        _createRegexError = _createRegexWriterError  if _bWriterError  else _createRegexDictError
    except:
        traceback.print_exc()


def _getRules (bParagraph):
    try:
        if not bParagraph:
            return _rules.lSentenceRules
        return _rules.lParagraphRules
    except:
        _loadRules()
    if not bParagraph:
        return _rules.lSentenceRules
    return _rules.lParagraphRules


def _loadRules ():
    from . import gc_rules
    global _rules
    _rules = gc_rules
    # compile rules regex
    for sOption, lRuleGroup in chain(_rules.lParagraphRules, _rules.lSentenceRules):
        if sOption != "@@@@":
            for aRule in lRuleGroup:
                try:
                    aRule[0] = re.compile(aRule[0])
                except:
                    echo("Bad regular expression in # " + str(aRule[2]))
                    aRule[0] = "(?i)<Grammalecte>"


#### Parsing

_zEndOfSentence = re.compile(r'([.?!:;…][ .?!… »”")]*|.$)')
_zBeginOfParagraph = re.compile(r"^\W*")
_zEndOfParagraph = re.compile(r"\W*$")

def _getSentenceBoundaries (sText):
    iStart = _zBeginOfParagraph.match(sText).end()
    for m in _zEndOfSentence.finditer(sText):
        yield (iStart, m.end())
        iStart = m.end()


def parse (sText, sCountry="${country_default}", bDebug=False, dOptions=None, bContext=False):
    "analyses the paragraph sText and returns list of errors"
    #sText = unicodedata.normalize("NFC", sText)
    aErrors = None
    sAlt = sText
    sRealText = sText
    dDA = {}        # Disambiguisator. Key = position; value = list of morphologies
    dPriority = {}  # Key = position; value = priority
    dOpt = _dOptions  if not dOptions  else dOptions
    bShowRuleId = option('idrule')

    # parse paragraph
    try:
        sNew, aErrors = _proofread(sText, sAlt, 0, True, dDA, dPriority, sCountry, dOpt, bDebug, bContext)
        sNew, aErrors = _proofread(None, sText, sRealText, 0, True, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
        if sNew:
            sText = sNew
    except:
        raise

    # cleanup
    if " " in sText:
        sText = sText.replace(" ", ' ') # nbsp
    if " " in sText:
        sText = sText.replace(" ", ' ') # nnbsp
    if "'" in sText:
        sText = sText.replace("'", "’")
    if "‑" in sText:
        sText = sText.replace("‑", "-") # nobreakdash

    # parse sentences
    for iStart, iEnd in _getSentenceBoundaries(sText):
        if 4 < (iEnd - iStart) < 2000:
            dDA.clear()
            try:
                _, errs = _proofread(sText[iStart:iEnd], sAlt[iStart:iEnd], iStart, False, dDA, dPriority, sCountry, dOpt, bDebug, bContext)
                oSentence = TokenSentence(sText[iStart:iEnd], sRealText[iStart:iEnd], iStart)
                _, errs = _proofread(oSentence, sText[iStart:iEnd], sRealText[iStart:iEnd], iStart, False, dPriority, sCountry, dOpt, bShowRuleId, bDebug, bContext)
                aErrors.update(errs)
            except:
                raise
    return aErrors.values() # this is a view (iterable)


def _getSentenceBoundaries (sText):
    iStart = _zBeginOfParagraph.match(sText).end()
    for m in _zEndOfSentence.finditer(sText):
        yield (iStart, m.end())
        iStart = m.end()

def _proofread (oSentence, s, sx, nOffset, bParagraph, dPriority, sCountry, dOptions, bShowRuleId, bDebug, bContext):

def _proofread (s, sx, nOffset, bParagraph, dDA, dPriority, sCountry, dOptions, bDebug, bContext):
    dErrs = {}
    bChange = False
    bIdRule = option('idrule')

    bParagraphChange = False
    bSentenceChange = False
    dTokenPos = oSentence.dTokenPos if oSentence else {}
    for sOption, lRuleGroup in _getRules(bParagraph):
        if sOption == "@@@@":
            # graph rules
            if not bParagraph and bSentenceChange:
                oSentence.update(s)
                bSentenceChange = False
            for sGraphName, sLineId in lRuleGroup:
                if bDebug:
                    print("\n>>>> GRAPH:", sGraphName, sLineId)
                bParagraphChange, s = oSentence.parse(dAllGraph[sGraphName], dPriority, sCountry, dOptions, bShowRuleId, bDebug, bContext)
                dErrs.update(oSentence.dError)
        if not sOption or dOptions.get(sOption, False):
        elif not sOption or dOptions.get(sOption, False):
            # regex rules
            for zRegex, bUppercase, sLineId, sRuleId, nPriority, lActions in lRuleGroup:
                if sRuleId not in _aIgnoredRules:
                    for m in zRegex.finditer(s):
                        bCondMemo = None
                        for sFuncCond, cActionType, sWhat, *eAct in lActions:
                            # action in lActions: [ condition, action type, replacement/suggestion/action[, iGroup[, message, URL]] ]
                            try:
                                bCondMemo = not sFuncCond or globals()[sFuncCond](s, sx, m, dDA, sCountry, bCondMemo)
                                bCondMemo = not sFuncCond or globals()[sFuncCond](s, sx, m, dTokenPos, sCountry, bCondMemo)
                                if bCondMemo:
                                    if cActionType == "-":
                                        # grammar error
                                        nErrorStart = nOffset + m.start(eAct[0])
                                        if nErrorStart not in dErrs or nPriority > dPriority[nErrorStart]:
                                            dErrs[nErrorStart] = _createError(s, sx, sWhat, nOffset, m, eAct[0], sLineId, sRuleId, bUppercase, eAct[1], eAct[2], bIdRule, sOption, bContext)
                                        if nErrorStart not in dErrs or nPriority > dPriority.get(nErrorStart, -1):
                                            dErrs[nErrorStart] = _createRegexError(s, sx, sWhat, nOffset, m, eAct[0], sLineId, sRuleId, bUppercase, eAct[1], eAct[2], bShowRuleId, sOption, bContext)
                                            dPriority[nErrorStart] = nPriority
                                    elif cActionType == "~":
                                        # text processor
                                        s = _rewrite(s, sWhat, eAct[0], m, bUppercase)
                                        bChange = True
                                        bParagraphChange = True
                                        bSentenceChange = True
                                        if bDebug:
                                            echo("~ " + s + "  -- " + m.group(eAct[0]) + "  # " + sLineId)
                                    elif cActionType == "=":
                                        # disambiguation
                                        if not bParagraph:
                                        globals()[sWhat](s, m, dDA)
                                        if bDebug:
                                            echo("= " + m.group(0) + "  # " + sLineId + "\nDA: " + str(dDA))
                                            globals()[sWhat](s, m, dTokenPos)
                                            if bDebug:
                                                echo("= " + m.group(0) + "  # " + sLineId)
                                    elif cActionType == ">":
                                        # we do nothing, this test is just a condition to apply all following actions
                                        pass
                                    else:
                                        echo("# error: unknown action at " + sLineId)
                                elif cActionType == ">":
                                    break
                            except Exception as e:
                                raise Exception(str(e), "# " + sLineId + " # " + sRuleId)
    if bChange:
    if bParagraphChange:
        return (s, dErrs)
    return (False, dErrs)


def _createWriterError (s, sx, sRepl, nOffset, m, iGroup, sLineId, sRuleId, bUppercase, sMsg, sURL, bIdRule, sOption, bContext):
def _createRegexWriterError (s, sx, sRepl, nOffset, m, iGroup, sLineId, sRuleId, bUppercase, sMsg, sURL, bShowRuleId, sOption, bContext):
    "error for Writer (LO/OO)"
    xErr = SingleProofreadingError()
    #xErr = uno.createUnoStruct( "com.sun.star.linguistic2.SingleProofreadingError" )
    xErr.nErrorStart = nOffset + m.start(iGroup)
    xErr.nErrorLength = m.end(iGroup) - m.start(iGroup)
    xErr.nErrorType = PROOFREADING
    xErr.aRuleIdentifier = sRuleId

        xErr.aSuggestions = ()
    else:
        if bUppercase and m.group(iGroup)[0:1].isupper():
            xErr.aSuggestions = tuple(map(str.capitalize, m.expand(sRepl).split("|")))
        else:
            xErr.aSuggestions = tuple(m.expand(sRepl).split("|"))
    # Message
    if sMsg[0:1] == "=":
        sMessage = globals()[sMsg[1:]](s, m)
    sMessage = globals()[sMsg[1:]](s, m)  if sMsg[0:1] == "="  else  m.expand(sMsg)
    else:
        sMessage = m.expand(sMsg)
    xErr.aShortComment = sMessage   # sMessage.split("|")[0]     # in context menu
    xErr.aFullComment = sMessage   # sMessage.split("|")[-1]    # in dialog
    if bIdRule:
    if bShowRuleId:
        xErr.aShortComment += "  # " + sLineId + " # " + sRuleId
    # URL
    if sURL:
        p = PropertyValue()
        p.Name = "FullCommentURL"
        p.Value = sURL
        xErr.aProperties = (p,)
        xProperty = PropertyValue()
        xProperty.Name = "FullCommentURL"
        xProperty.Value = sURL
        xErr.aProperties = (xProperty,)
    else:
        xErr.aProperties = ()
    return xErr


def _createDictError (s, sx, sRepl, nOffset, m, iGroup, sLineId, sRuleId, bUppercase, sMsg, sURL, bIdRule, sOption, bContext):
def _createRegexDictError (s, sx, sRepl, nOffset, m, iGroup, sLineId, sRuleId, bUppercase, sMsg, sURL, bShowRuleId, sOption, bContext):
    "error as a dictionary"
    dErr = {}
    dErr["nStart"] = nOffset + m.start(iGroup)
    dErr["nEnd"] = nOffset + m.end(iGroup)
    dErr["sLineId"] = sLineId
    dErr["sRuleId"] = sRuleId
    dErr["sType"] = sOption  if sOption  else "notype"
    # suggestions
    if sRepl[0:1] == "=":
        sugg = globals()[sRepl[1:]](s, m)
        if sugg:
            if bUppercase and m.group(iGroup)[0:1].isupper():
                dErr["aSuggestions"] = list(map(str.capitalize, sugg.split("|")))
            else:
                dErr["aSuggestions"] = sugg.split("|")
        else:
            dErr["aSuggestions"] = ()
            dErr["aSuggestions"] = []
    elif sRepl == "_":
        dErr["aSuggestions"] = ()
        dErr["aSuggestions"] = []
    else:
        if bUppercase and m.group(iGroup)[0:1].isupper():
            dErr["aSuggestions"] = list(map(str.capitalize, m.expand(sRepl).split("|")))
        else:
            dErr["aSuggestions"] = m.expand(sRepl).split("|")
    # Message
    if sMsg[0:1] == "=":
        sMessage = globals()[sMsg[1:]](s, m)
    else:
        sMessage = m.expand(sMsg)
    dErr["sMessage"] = sMessage
    if bIdRule:
    dErr["sMessage"] = globals()[sMsg[1:]](s, m)  if sMsg[0:1] == "="  else  m.expand(sMsg)
    if bShowRuleId:
        dErr["sMessage"] += "  # " + sLineId + " # " + sRuleId
    # URL
    dErr["URL"] = sURL  if sURL  else ""
    # Context
    if bContext:
        dErr['sUnderlined'] = sx[m.start(iGroup):m.end(iGroup)]
        dErr['sBefore'] = sx[max(0,m.start(iGroup)-80):m.start(iGroup)]
        dErr['sAfter'] = sx[m.end(iGroup):m.end(iGroup)+80]
    return dErr


def _rewrite (s, sRepl, iGroup, m, bUppercase):
    "text processor: write sRepl in s at iGroup position"
def _rewrite (sSentence, sRepl, iGroup, m, bUppercase):
    "text processor: write <sRepl> in <sSentence> at <iGroup> position"
    nLen = m.end(iGroup) - m.start(iGroup)
    if sRepl == "*":
        sNew = " " * nLen
    elif sRepl == ">" or sRepl == "_" or sRepl == "~":
    elif sRepl == "_":
        sNew = sRepl + " " * (nLen-1)
    elif sRepl == "@":
        sNew = "@" * nLen
    elif sRepl[0:1] == "=":
        sNew = globals()[sRepl[1:]](s, m)
        sNew = globals()[sRepl[1:]](sSentence, m)
        sNew = sNew + " " * (nLen-len(sNew))
        if bUppercase and m.group(iGroup)[0:1].isupper():
            sNew = sNew.capitalize()
    else:
        sNew = m.expand(sRepl)
        sNew = sNew + " " * (nLen-len(sNew))
    return s[0:m.start(iGroup)] + sNew + s[m.end(iGroup):]
    return sSentence[0:m.start(iGroup)] + sNew + sSentence[m.end(iGroup):]


def ignoreRule (sRuleId):
    "disable rule <sRuleId>"
    _aIgnoredRules.add(sRuleId)


def resetIgnoreRules ():
    "clear all ignored rules"
    _aIgnoredRules.clear()


def reactivateRule (sRuleId):
    "(re)activate rule <sRuleId>"
    _aIgnoredRules.discard(sRuleId)


def listRules (sFilter=None):
    "generator: returns typle (sOption, sLineId, sRuleId)"
    if sFilter:
        try:
            zFilter = re.compile(sFilter)
        except:
            echo("# Error. List rules: wrong regex.")
            sFilter = None
    for sOption, lRuleGroup in chain(_getRules(True), _getRules(False)):
        if sOption != "@@@@":
        for _, _, sLineId, sRuleId, _, _ in lRuleGroup:
            if not sFilter or zFilter.search(sRuleId):
                yield (sOption, sLineId, sRuleId)
            for _, _, sLineId, sRuleId, _, _ in lRuleGroup:
                if not sFilter or zFilter.search(sRuleId):
                    yield (sOption, sLineId, sRuleId)


def displayRules (sFilter=None):
    "display the name of rules, with the filter <sFilter>"
    echo("List of rules. Filter: << " + str(sFilter) + " >>")
    for sOption, sLineId, sRuleId in listRules(sFilter):
        echo("{:<10} {:<10} {}".format(sOption, sLineId, sRuleId))


#### init

try:
    # LibreOffice / OpenOffice
    from com.sun.star.linguistic2 import SingleProofreadingError
    from com.sun.star.text.TextMarkupType import PROOFREADING
    from com.sun.star.beans import PropertyValue
    #import lightproof_handler_${implname} as opt
    _createError = _createWriterError
except ImportError:
    _createError = _createDictError


def load (sContext="Python"):
    global _oSpellChecker
    global _sAppContext
    global _dOptions
    try:
        _oSpellChecker = SpellChecker("${lang}", "${dic_main_filename_py}", "${dic_extended_filename_py}", "${dic_community_filename_py}", "${dic_personal_filename_py}")
        _sAppContext = sContext
        _dOptions = dict(gc_options.getOptions(sContext))   # duplication necessary, to be able to reset to default
    except:
        traceback.print_exc()


def setOption (sOpt, bVal):
    "set option <sOpt> with <bVal> if it exists"
    if sOpt in _dOptions:
        _dOptions[sOpt] = bVal


def setOptions (dOpt):
    "update the dictionary of options with <dOpt>"
    for sKey, bVal in dOpt.items():
        if sKey in _dOptions:
            _dOptions[sKey] = bVal


def getOptions ():
    "return the dictionary of current options"
    return _dOptions


def getDefaultOptions ():
    "return the dictionary of default options"
    return dict(gc_options.getOptions(_sAppContext))


def getOptionsLabels (sLang):
    "return options labels"
    return gc_options.getUI(sLang)


def displayOptions (sLang):
    "display the list of grammar checking options"
    echo("List of options")
    echo("\n".join( [ k+":\t"+str(v)+"\t"+gc_options.getUI(sLang).get(k, ("?", ""))[0]  for k, v  in sorted(_dOptions.items()) ] ))
    echo("")


def resetOptions ():
    "set options to default values"
    global _dOptions
    _dOptions = dict(gc_options.getOptions(_sAppContext))


def getSpellChecker ():
    "return the spellchecker object"
    return _oSpellChecker


def _getRules (bParagraph):
    try:
        if not bParagraph:
            return _rules.lSentenceRules
        return _rules.lParagraphRules
    except:
        _loadRules()
    if not bParagraph:
        return _rules.lSentenceRules
    return _rules.lParagraphRules


def _loadRules ():
    from . import gc_rules
    global _rules
    _rules = gc_rules
    # compile rules regex
    for lRuleGroup in chain(_rules.lParagraphRules, _rules.lSentenceRules):
        for rule in lRuleGroup[1]:
            try:
                rule[0] = re.compile(rule[0])
            except:
                echo("Bad regular expression in # " + str(rule[2]))
                rule[0] = "(?i)<Grammalecte>"


def _getPath ():
    return os.path.join(os.path.dirname(sys.modules[__name__].__file__), __name__ + ".py")



#### common functions

# common regexes
_zEndOfSentence = re.compile('([.?!:;…][ .?!… »”")]*|.$)')
_zBeginOfParagraph = re.compile("^\W*")
_zEndOfParagraph = re.compile("\W*$")
_zNextWord = re.compile(" +(\w[\w-]*)")
_zPrevWord = re.compile("(\w[\w-]*) +$")


def option (sOpt):
    "return True if option sOpt is active"
    "return True if option <sOpt> is active"
    return _dOptions.get(sOpt, False)


def displayInfo (dDA, tWord):
def displayInfo (dTokenPos, tWord):
    "for debugging: retrieve info of word"
    if not tWord:
        echo("> nothing to find")
        return True
    if tWord[1] not in _dAnalyses and not _storeMorphFromFSA(tWord[1]):
        echo("> not in FSA")
    lMorph = _oSpellChecker.getMorph(tWord[1])
    if not lMorph:
        echo("> not in dictionary")
        return True
    if tWord[0] in dDA:
        echo("DA: " + str(dDA[tWord[0]]))
    echo("FSA: " + str(_dAnalyses[tWord[1]]))
    if tWord[0] in dTokenPos and "lMorph" in dTokenPos[tWord[0]]:
        echo("DA: " + str(dTokenPos[tWord[0]]["lMorph"]))
    echo("FSA: " + str(lMorph))
    return True


def _storeMorphFromFSA (sWord):
    "retrieves morphologies list from _oSpellChecker -> _dAnalyses"
    global _dAnalyses
    _dAnalyses[sWord] = _oSpellChecker.getMorph(sWord)
    return True  if _dAnalyses[sWord]  else False


def morph (dDA, tWord, sPattern, bStrict=True, bNoWord=False):
def morph (dTokenPos, tWord, sPattern, bStrict=True, bNoWord=False):
    "analyse a tuple (position, word), return True if sPattern in morphologies (disambiguation on)"
    if not tWord:
        return bNoWord
    if tWord[1] not in _dAnalyses and not _storeMorphFromFSA(tWord[1]):
        return False
    lMorph = dDA[tWord[0]]  if tWord[0] in dDA  else _dAnalyses[tWord[1]]
    lMorph = dTokenPos[tWord[0]]["lMorph"]  if tWord[0] in dTokenPos and "lMorph" in dTokenPos[tWord[0]]  else _oSpellChecker.getMorph(tWord[1])
    if not lMorph:
        return False
    p = re.compile(sPattern)
    zPattern = re.compile(sPattern)
    if bStrict:
        return all(p.search(s)  for s in lMorph)
    return any(p.search(s)  for s in lMorph)
        return all(zPattern.search(s)  for s in lMorph)
    return any(zPattern.search(s)  for s in lMorph)


def morphex (dDA, tWord, sPattern, sNegPattern, bNoWord=False):
def morphex (dTokenPos, tWord, sPattern, sNegPattern, bNoWord=False):
    "analyse a tuple (position, word), returns True if not sNegPattern in word morphologies and sPattern in word morphologies (disambiguation on)"
    if not tWord:
        return bNoWord
    lMorph = dTokenPos[tWord[0]]["lMorph"]  if tWord[0] in dTokenPos and "lMorph" in dTokenPos[tWord[0]]  else _oSpellChecker.getMorph(tWord[1])
    if tWord[1] not in _dAnalyses and not _storeMorphFromFSA(tWord[1]):
    if not lMorph:
        return False
    lMorph = dDA[tWord[0]]  if tWord[0] in dDA  else _dAnalyses[tWord[1]]
    # check negative condition
    np = re.compile(sNegPattern)
    if any(np.search(s)  for s in lMorph):
    zNegPattern = re.compile(sNegPattern)
    if any(zNegPattern.search(s)  for s in lMorph):
        return False
    # search sPattern
    p = re.compile(sPattern)
    return any(p.search(s)  for s in lMorph)
    zPattern = re.compile(sPattern)
    return any(zPattern.search(s)  for s in lMorph)


def analyse (sWord, sPattern, bStrict=True):
    "analyse a word, return True if sPattern in morphologies (disambiguation off)"
    if sWord not in _dAnalyses and not _storeMorphFromFSA(sWord):
    lMorph = _oSpellChecker.getMorph(sWord)
        return False
    if not _dAnalyses[sWord]:
    if not lMorph:
        return False
    p = re.compile(sPattern)
    zPattern = re.compile(sPattern)
    if bStrict:
        return all(p.search(s)  for s in _dAnalyses[sWord])
    return any(p.search(s)  for s in _dAnalyses[sWord])
        return all(zPattern.search(s)  for s in lMorph)
    return any(zPattern.search(s)  for s in lMorph)


def analysex (sWord, sPattern, sNegPattern):
    "analyse a word, returns True if not sNegPattern in word morphologies and sPattern in word morphologies (disambiguation off)"
    if sWord not in _dAnalyses and not _storeMorphFromFSA(sWord):
    lMorph = _oSpellChecker.getMorph(sWord)
    if not lMorph:
        return False
    # check negative condition
    np = re.compile(sNegPattern)
    if any(np.search(s)  for s in _dAnalyses[sWord]):
    zNegPattern = re.compile(sNegPattern)
    if any(zNegPattern.search(s)  for s in lMorph):
        return False
    # search sPattern
    p = re.compile(sPattern)
    return any(p.search(s)  for s in _dAnalyses[sWord])
    zPattern = re.compile(sPattern)
    return any(zPattern.search(s)  for s in lMorph)


def stem (sWord):
    "returns a list of sWord's stems"
    if not sWord:
        return []
    if sWord not in _dAnalyses and not _storeMorphFromFSA(sWord):
        return []
    return [ s[1:s.find(" ")]  for s in _dAnalyses[sWord] ]


## functions to get text outside pattern scope

# warning: check compile_rules.py to understand how it works

_zNextWord = re.compile(r" +(\w[\w-]*)")
_zPrevWord = re.compile(r"(\w[\w-]*) +$")

def nextword (s, iStart, n):
    "get the nth word of the input string or empty string"
    m = re.match("(?: +[\\w%-]+){" + str(n-1) + "} +([\\w%-]+)", s[iStart:])
    if not m:
        return None
    return (iStart+m.start(1), m.group(1))

    if sNegPattern and re.search(sNegPattern, s):
        return False
    if re.search(sPattern, s):
        return True
    return False


def look_chk1 (dDA, s, nOffset, sPattern, sPatternGroup1, sNegPatternGroup1=None):
def look_chk1 (dTokenPos, s, nOffset, sPattern, sPatternGroup1, sNegPatternGroup1=None):
    "returns True if s has pattern sPattern and m.group(1) has pattern sPatternGroup1"
    m = re.search(sPattern, s)
    if not m:
        return False
    try:
        sWord = m.group(1)
        nPos = m.start(1) + nOffset
    except:
        return False
    if sNegPatternGroup1:
        return morphex(dDA, (nPos, sWord), sPatternGroup1, sNegPatternGroup1)
    return morph(dDA, (nPos, sWord), sPatternGroup1, False)
        return morphex(dTokenPos, (nPos, sWord), sPatternGroup1, sNegPatternGroup1)
    return morph(dTokenPos, (nPos, sWord), sPatternGroup1, False)


#### Disambiguator

def select (dTokenPos, nPos, sWord, sPattern, lDefault=None):
    "Disambiguation: select morphologies of <sWord> matching <sPattern>"
    if not sWord:
        return True
    if nPos not in dTokenPos:
        print("Error. There should be a token at this position: ", nPos)
        return True
    lMorph = _oSpellChecker.getMorph(sWord)
    if not lMorph or len(lMorph) == 1:
        return True
    lSelect = [ sMorph  for sMorph in lMorph  if re.search(sPattern, sMorph) ]
    if lSelect:
        if len(lSelect) != len(lMorph):
            dTokenPos[nPos]["lMorph"] = lSelect
    elif lDefault:
        dTokenPos[nPos]["lMorph"] = lDefault
    return True


def exclude (dTokenPos, nPos, sWord, sPattern, lDefault=None):
    "Disambiguation: exclude morphologies of <sWord> matching <sPattern>"
    if not sWord:
        return True
    if nPos not in dTokenPos:
        print("Error. There should be a token at this position: ", nPos)
        return True
    lMorph = _oSpellChecker.getMorph(sWord)
    if not lMorph or len(lMorph) == 1:
        return True
    lSelect = [ sMorph  for sMorph in lMorph  if not re.search(sPattern, sMorph) ]
    if lSelect:
        if len(lSelect) != len(lMorph):
            dTokenPos[nPos]["lMorph"] = lSelect
    elif lDefault:
        dTokenPos[nPos]["lMorph"] = lDefault
    return True


def define (dTokenPos, nPos, lMorph):
    "Disambiguation: set morphologies of token at <nPos> with <lMorph>"
    if nPos not in dTokenPos:
        print("Error. There should be a token at this position: ", nPos)
        return True
    dTokenPos[nPos]["lMorph"] = lMorph
    return True




#### TOKEN SENTENCE CHECKER

class TokenSentence:
    "Text parser"

    def __init__ (self, sSentence, sSentence0, nOffset):
        self.sSentence = sSentence
        self.sSentence0 = sSentence0
        self.nOffsetWithinParagraph = nOffset
        self.lToken = list(_oTokenizer.genTokens(sSentence, True))
        self.dTokenPos = { dToken["nStart"]: dToken  for dToken in self.lToken }
        self.dTags = {}
        self.dError = {}
        self.createError = self._createWriterError  if _bWriterError  else self._createDictError

    def update (self, sSentence):
        "update <sSentence> and retokenize"
        self.sSentence = sSentence
        self.lToken = list(_oTokenizer.genTokens(sSentence, True))

    def _getNextMatchingNodes (self, dToken, dGraph, dNode, bDebug=False):
        "generator: return nodes where <dToken> “values” match <dNode> arcs"
        # token value
        if dToken["sValue"] in dNode:
            if bDebug:
                print("  MATCH:", dToken["sValue"])
            yield dGraph[dNode[dToken["sValue"]]]
        if dToken["sValue"][0:2].istitle(): # we test only 2 first chars, to make valid words such as "Laissez-les", "Passe-partout".
            sValue = dToken["sValue"].lower()
            if sValue in dNode:
                if bDebug:
                    print("  MATCH:", sValue)
                yield dGraph[dNode[sValue]]
        elif dToken["sValue"].isupper():
            sValue = dToken["sValue"].lower()
            if sValue in dNode:
                if bDebug:
                    print("  MATCH:", sValue)
                yield dGraph[dNode[sValue]]
            sValue = dToken["sValue"].capitalize()
            if sValue in dNode:
                if bDebug:
                    print("  MATCH:", sValue)
                yield dGraph[dNode[sValue]]
        # token lemmas
        if "<lemmas>" in dNode:
            for sLemma in _oSpellChecker.getLemma(dToken["sValue"]):
                if sLemma in dNode["<lemmas>"]:
                    if bDebug:
                        print("  MATCH: >" + sLemma)
                    yield dGraph[dNode["<lemmas>"][sLemma]]
        # regex value arcs
        if "<re_value>" in dNode:
            for sRegex in dNode["<re_value>"]:
                if "¬" not in sRegex:
                    # no anti-pattern
                    if re.search(sRegex, dToken["sValue"]):
                        if bDebug:
                            print("  MATCH: ~" + sRegex)
                        yield dGraph[dNode["<re_value>"][sRegex]]
                else:
                    # there is an anti-pattern
                    sPattern, sNegPattern = sRegex.split("¬", 1)
                    if sNegPattern and re.search(sNegPattern, dToken["sValue"]):
                        continue
                    if not sPattern or re.search(sPattern, dToken["sValue"]):
                        if bDebug:
                            print("  MATCH: ~" + sRegex)
                        yield dGraph[dNode["<re_value>"][sRegex]]
        # regex morph arcs
        if "<re_morph>" in dNode:
            for sRegex in dNode["<re_morph>"]:
                if "¬" not in sRegex:
                    # no anti-pattern
                    if any(re.search(sRegex, sMorph)  for sMorph in _oSpellChecker.getMorph(dToken["sValue"])):
                        if bDebug:
                            print("  MATCH: @" + sRegex)
                        yield dGraph[dNode["<re_morph>"][sRegex]]
                else:
                    # there is an anti-pattern
                    sPattern, sNegPattern = sRegex.split("¬", 1)
                    if sNegPattern == "*":
                        # all morphologies must match with <sPattern>
                        if sPattern and all(re.search(sPattern, sMorph)  for sMorph in _oSpellChecker.getMorph(dToken["sValue"])):
                            if bDebug:
                                print("  MATCH: @" + sRegex)
                            yield dGraph[dNode["<re_morph>"][sRegex]]
                    else:
                        if sNegPattern and any(re.search(sNegPattern, sMorph)  for sMorph in _oSpellChecker.getMorph(dToken["sValue"])):
                            continue
                        if not sPattern or any(re.search(sPattern, sMorph)  for sMorph in _oSpellChecker.getMorph(dToken["sValue"])):
                            if bDebug:
                                print("  MATCH: @" + sRegex)
                            yield dGraph[dNode["<re_morph>"][sRegex]]
        # meta arc (for token type)
        if "<meta>" in dNode:
            for sMeta in dNode["<meta>"]:
                # not regex here, we just search if <dNode["sType"]> exists within <sMeta>
                if sMeta == "*":
                    if bDebug:
                        print("  MATCH: *" + sMeta)
                    yield dGraph[dNode["<meta>"]["*"]]
                elif "¬" in sMeta:
                    if dNode["sType"] not in sMeta:
                        if bDebug:
                            print("  MATCH: *" + sMeta)
                        yield dGraph[dNode["<meta>"][sMeta]]
                elif dNode["sType"] in sMeta:
                    if bDebug:
                        print("  MATCH: *" + sMeta)
                    yield dGraph[dNode["<meta>"][sMeta]]


    def parse (self, dGraph, dPriority, sCountry="${country_default}", dOptions=None, bShowRuleId=False, bDebug=False, bContext=False):
        "parse tokens from the text and execute actions encountered"
        self.dError = {}
        dPriority = {}  # Key = position; value = priority
        dOpt = _dOptions  if not dOptions  else dOptions
        lPointer = []
        bTagAndRewrite = False
        for dToken in self.lToken:
            if bDebug:
                print("TOKEN:", dToken["sValue"])
            # check arcs for each existing pointer
            lNextPointer = []
            for dPointer in lPointer:
                for dNode in self._getNextMatchingNodes(dToken, dGraph, dPointer["dNode"], bDebug):
                    lNextPointer.append({"iToken": dPointer["iToken"], "dNode": dNode})
            lPointer = lNextPointer
            # check arcs of first nodes
            for dNode in self._getNextMatchingNodes(dToken, dGraph, dGraph[0], bDebug):
                lPointer.append({"iToken": dToken["i"], "dNode": dNode})
            # check if there is rules to check for each pointer
            for dPointer in lPointer:
                #if bDebug:
                #    print("+", dPointer)
                if "<rules>" in dPointer["dNode"]:
                    bChange, dErr = self._executeActions(dGraph, dPointer["dNode"]["<rules>"], dPointer["iToken"]-1, dToken["i"], dPriority, dOpt, sCountry, bShowRuleId, bDebug, bContext)
                    self.dError.update(dErr)
                    if bChange:
                        bTagAndRewrite = True
        if bTagAndRewrite:
            self.rewrite(bDebug)
        return (bTagAndRewrite, self.sSentence)

    def _executeActions (self, dGraph, dNode, nTokenOffset, nLastToken, dPriority, dOptions, sCountry, bShowRuleId, bDebug, bContext):
        "execute actions found in the DARG"
        dError = {}
        bChange = False
        for sLineId, nextNodeKey in dNode.items():
            bCondMemo = None
            for sRuleId in dGraph[nextNodeKey]:
                try:
                    if bDebug:
                        print("ACTION:", sRuleId)
                        print(dRule[sRuleId])
                    sOption, sFuncCond, cActionType, sWhat, *eAct = dRule[sRuleId]
                    # Suggestion    [ option, condition, "-", replacement/suggestion/action, iTokenStart, iTokenEnd, nPriority, message, URL ]
                    # TextProcessor [ option, condition, "~", replacement/suggestion/action, iTokenStart, iTokenEnd ]
                    # Disambiguator [ option, condition, "=", replacement/suggestion/action ]
                    # Sentence Tag  [ option, condition, "/", replacement/suggestion/action, iTokenStart, iTokenEnd ]
                    # Test          [ option, condition, ">", "" ]
                    if not sOption or dOptions.get(sOption, False):
                        bCondMemo = not sFuncCond or globals()[sFuncCond](self.lToken, nTokenOffset, nLastToken, sCountry, bCondMemo, self.dTags, self.sSentence, self.sSentence0)
                        if bCondMemo:
                            if cActionType == "-":
                                # grammar error
                                nTokenErrorStart = nTokenOffset + eAct[0]
                                if "bImmune" not in self.lToken[nTokenErrorStart]:
                                    nTokenErrorEnd = (nTokenOffset + eAct[1])  if eAct[1]  else nLastToken
                                    nErrorStart = self.nOffsetWithinParagraph + self.lToken[nTokenErrorStart]["nStart"]
                                    nErrorEnd = self.nOffsetWithinParagraph + self.lToken[nTokenErrorEnd]["nEnd"]
                                    if nErrorStart not in dError or eAct[2] > dPriority.get(nErrorStart, -1):
                                        dError[nErrorStart] = self.createError(sWhat, nTokenOffset, nTokenErrorStart, nErrorStart, nErrorEnd, sLineId, sRuleId, True, eAct[3], eAct[4], bShowRuleId, "notype", bContext)
                                        dPriority[nErrorStart] = eAct[2]
                                        if bDebug:
                                            print("-", sRuleId, dError[nErrorStart])
                            elif cActionType == "~":
                                # text processor
                                nEndToken = (nTokenOffset + eAct[1])  if eAct[1]  else nLastToken
                                self._tagAndPrepareTokenForRewriting(sWhat, nTokenOffset + eAct[0], nEndToken, bDebug)
                                if bDebug:
                                    print("~", sRuleId)
                                bChange = True
                            elif cActionType == "=":
                                # disambiguation
                                globals()[sWhat](self.lToken, nTokenOffset)
                                if bDebug:
                                    print("=", sRuleId)
                            elif cActionType == ">":
                                # we do nothing, this test is just a condition to apply all following actions
                                if bDebug:
                                    print(">", sRuleId)
                                pass
                            elif cActionType == "/":
                                # tags
                                nTokenTag = nTokenOffset + eAct[0]
                                if sWhat not in self.dTags:
                                    self.dTags[sWhat] = (nTokenTag, nTokenTag)
                                elif nTokenTag > self.dTags[sWhat][1]:
                                    self.dTags[sWhat] = (self.dTags[sWhat][0], nTokenTag)
                                if bDebug:
                                    print("/", sRuleId)
                            else:
                                print("# error: unknown action at " + sLineId)
                        elif cActionType == ">":
                            if bDebug:
                                print(">!", sRuleId)
                            break
                except Exception as e:
                    raise Exception(str(e), sLineId, sRuleId, self.sSentence)
        return bChange, dError

    def _createWriterError (self, sSugg, nTokenOffset, iFirstToken, nStart, nEnd, sLineId, sRuleId, bUppercase, sMsg, sURL, bShowRuleId, sOption, bContext):
        "error for Writer (LO/OO)"
        xErr = SingleProofreadingError()
        #xErr = uno.createUnoStruct( "com.sun.star.linguistic2.SingleProofreadingError" )
        xErr.nErrorStart = nStart
        xErr.nErrorLength = nEnd - nStart
        xErr.nErrorType = PROOFREADING
        xErr.aRuleIdentifier = sRuleId
        # suggestions
        if sSugg[0:1] == "=":
            sSugg = globals()[sSugg[1:]](self.lToken, nTokenOffset)
            if sSugg:
                if bUppercase and self.lToken[iFirstToken]["sValue"][0:1].isupper():
                    xErr.aSuggestions = tuple(map(str.capitalize, sSugg.split("|")))
                else:
                    xErr.aSuggestions = tuple(sSugg.split("|"))
            else:
                xErr.aSuggestions = ()
        elif sSugg == "_":
            xErr.aSuggestions = ()
        else:
            if bUppercase and self.lToken[iFirstToken]["sValue"][0:1].isupper():
                xErr.aSuggestions = tuple(map(str.capitalize, self._expand(sSugg, nTokenOffset).split("|")))
            else:
                xErr.aSuggestions = tuple(self._expand(sSugg, nTokenOffset).split("|"))
        # Message
        sMessage = globals()[sMsg[1:]](self.lToken)  if sMsg[0:1] == "="  else self._expand(sMsg, nTokenOffset)
        xErr.aShortComment = sMessage   # sMessage.split("|")[0]     # in context menu
        xErr.aFullComment = sMessage   # sMessage.split("|")[-1]    # in dialog
        if bShowRuleId:
            xErr.aShortComment += "  " + sLineId + " # " + sRuleId
        # URL
        if sURL:
            xProperty = PropertyValue()
            xProperty.Name = "FullCommentURL"
            xProperty.Value = sURL
            xErr.aProperties = (xProperty,)
        else:
            xErr.aProperties = ()
        return xErr

    def _createDictError (self, sSugg, nTokenOffset, iFirstToken, nStart, nEnd, sLineId, sRuleId, bUppercase, sMsg, sURL, bShowRuleId, sOption, bContext):
        "error as a dictionary"
        dErr = {}
        dErr["nStart"] = nStart
        dErr["nEnd"] = nEnd
        dErr["sLineId"] = sLineId
        dErr["sRuleId"] = sRuleId
        dErr["sType"] = sOption  if sOption  else "notype"
        # suggestions
        if sSugg[0:1] == "=":
            sSugg = globals()[sSugg[1:]](self.lToken, nTokenOffset)
            if sSugg:
                if bUppercase and self.lToken[iFirstToken]["sValue"][0:1].isupper():
                    dErr["aSuggestions"] = list(map(str.capitalize, sSugg.split("|")))
                else:
                    dErr["aSuggestions"] = sSugg.split("|")
            else:
                dErr["aSuggestions"] = []
        elif sSugg == "_":
            dErr["aSuggestions"] = []
        else:
            if bUppercase and self.lToken[iFirstToken]["sValue"][0:1].isupper():
                dErr["aSuggestions"] = list(map(str.capitalize, self._expand(sSugg, nTokenOffset).split("|")))
            else:
                dErr["aSuggestions"] = self._expand(sSugg, nTokenOffset).split("|")
        # Message
        dErr["sMessage"] = globals()[sMsg[1:]](self.lToken)  if sMsg[0:1] == "="  else self._expand(sMsg, nTokenOffset)
        if bShowRuleId:
            dErr["sMessage"] += "  " + sLineId + " # " + sRuleId
        # URL
        dErr["URL"] = sURL  if sURL  else ""
        # Context
        if bContext:
            dErr['sUnderlined'] = self.sSentence0[dErr["nStart"]:dErr["nEnd"]]
            dErr['sBefore'] = self.sSentence0[max(0,dErr["nStart"]-80):dErr["nStart"]]
            dErr['sAfter'] = self.sSentence0[dErr["nEnd"]:dErr["nEnd"]+80]
        return dErr

    def _expand (self, sMsg, nTokenOffset):
        #print("*", sMsg)
        for m in re.finditer(r"\\([0-9]+)", sMsg):
            sMsg = sMsg.replace(m.group(0), self.lToken[int(m.group(1))+nTokenOffset]["sValue"])
        #print(">", sMsg)
        return sMsg

    def _tagAndPrepareTokenForRewriting (self, sWhat, nTokenRewriteStart, nTokenRewriteEnd, bUppercase=True, bDebug=False):
        "text processor: rewrite tokens between <nTokenRewriteStart> and <nTokenRewriteEnd> position"
        if bDebug:
            print("REWRITING:", nTokenRewriteStart, nTokenRewriteEnd)
        if sWhat == "*":
            # purge text
            if nTokenRewriteEnd - nTokenRewriteStart == 0:
                self.lToken[nTokenRewriteStart]["bToRemove"] = True
            else:
                for i in range(nTokenRewriteStart, nTokenRewriteEnd+1):
                    self.lToken[i]["bToRemove"] = True
        elif sWhat == "_":
            # merge tokens
            self.lToken[nTokenRewriteStart]["nMergeUntil"] = nTokenRewriteEnd
        elif sWhat == "!":
            # immunity
            if nTokenRewriteEnd - nTokenRewriteStart == 0:
                self.lToken[nTokenRewriteStart]["bImmune"] = True
            else:
                for i in range(nTokenRewriteStart, nTokenRewriteEnd+1):
                    self.lToken[i]["bImmune"] = True
        else:
            if sWhat.startswith("="):
                sWhat = globals()[sWhat[1:]](self.lToken)
            bUppercase = bUppercase and self.lToken[nTokenRewriteStart]["sValue"][0:1].isupper()
            if nTokenRewriteEnd - nTokenRewriteStart == 0:
                sWhat = sWhat + " " * (len(self.lToken[nTokenRewriteStart]["sValue"])-len(sWhat))
                if bUppercase:
                    sWhat = sWhat[0:1].upper() + sWhat[1:]
                self.lToken[nTokenRewriteStart]["sNewValue"] = sWhat
            else:
                lTokenValue = sWhat.split("|")
                if len(lTokenValue) != (nTokenRewriteEnd - nTokenRewriteStart + 1):
                    print("Error. Text processor: number of replacements != number of tokens.")
                    return
                for i, sValue in zip(range(nTokenRewriteStart, nTokenRewriteEnd+1), lTokenValue):
                    if bUppercase:
                        sValue = sValue[0:1].upper() + sValue[1:]
                    self.lToken[i]["sNewValue"] = sValue

    def rewrite (self, bDebug=False):
        "rewrite the sentence, modify tokens, purge the token list"
        lNewToken = []
        nMergeUntil = 0
        dTokenMerger = None
        for dToken in self.lToken:
            bKeepToken = True
            if "bImmune" in dToken:
                nErrorStart = self.nOffsetWithinParagraph + dToken["nStart"]
                if nErrorStart in self.dError:
                    if bDebug:
                        print("immunity -> error removed:", self.dError[nErrorStart])
                    del self.dError[nErrorStart]
            if nMergeUntil and dToken["i"] <= nMergeUntil:
                dTokenMerger["sValue"] += " " * (dToken["nStart"] - dTokenMerger["nEnd"]) + dToken["sValue"]
                dTokenMerger["nEnd"] = dToken["nEnd"]
                if bDebug:
                    print("Merged token:", dTokenMerger["sValue"])
                bKeepToken = False
            if "nMergeUntil" in dToken:
                if dToken["i"] > nMergeUntil: # this token is not already merged with a previous token
                    dTokenMerger = dToken
                if dToken["nMergeUntil"] > nMergeUntil:
                    nMergeUntil = dToken["nMergeUntil"]
                del dToken["nMergeUntil"]
            elif "bToRemove" in dToken:
                # remove useless token
                self.sSentence = self.sSentence[:dToken["nStart"]] + " " * (dToken["nEnd"] - dToken["nStart"]) + self.sSentence[dToken["nEnd"]:]
                if bDebug:
                    print("removed:", dToken["sValue"])
                bKeepToken = False
            #
            if bKeepToken:
                lNewToken.append(dToken)
                if "sNewValue" in dToken:
                    # rewrite token and sentence
                    if bDebug:
                        print(dToken["sValue"], "->", dToken["sNewValue"])
                    dToken["sRealValue"] = dToken["sValue"]
                    dToken["sValue"] = dToken["sNewValue"]
                    nDiffLen = len(dToken["sRealValue"]) - len(dToken["sNewValue"])
                    sNewRepl = (dToken["sNewValue"] + " " * nDiffLen)  if nDiffLen >= 0  else dToken["sNewValue"][:len(dToken["sRealValue"])]
                    self.sSentence = self.sSentence[:dToken["nStart"]] + sNewRepl + self.sSentence[dToken["nEnd"]:]
                    del dToken["sNewValue"]
        if bDebug:
            print(self.sSentence)
        self.lToken.clear()
        self.lToken = lNewToken



#### Analyse tokens

def g_morph (dToken, sPattern, sNegPattern=""):
    "analyse a token, return True if <sNegPattern> not in morphologies and <sPattern> in morphologies"
    if "lMorph" in dToken:
        lMorph = dToken["lMorph"]
    else:
        lMorph = _oSpellChecker.getMorph(dToken["sValue"])
        if not lMorph:
            return False
    # check negative condition
    if sNegPattern:
        if sNegPattern == "*":
            # all morph must match sPattern
            zPattern = re.compile(sPattern)
            return all(zPattern.search(sMorph)  for sMorph in lMorph)
        else:
            zNegPattern = re.compile(sNegPattern)
            if any(zNegPattern.search(sMorph)  for sMorph in lMorph):
                return False
    # search sPattern
    zPattern = re.compile(sPattern)
    return any(zPattern.search(sMorph)  for sMorph in lMorph)


def g_analyse (dToken, sPattern, sNegPattern=""):
    "analyse a token, return True if <sNegPattern> not in morphologies and <sPattern> in morphologies (disambiguation off)"
    lMorph = _oSpellChecker.getMorph(dToken["sValue"])
    if not lMorph:
        return False
    # check negative condition
    if sNegPattern:
        if sNegPattern == "*":
            zPattern = re.compile(sPattern)
            return all(zPattern.search(sMorph)  for sMorph in lMorph)
        else:
            zNegPattern = re.compile(sNegPattern)
            if any(zNegPattern.search(sMorph)  for sMorph in lMorph):
                return False
    # search sPattern
    zPattern = re.compile(sPattern)
    return any(zPattern.search(sMorph)  for sMorph in lMorph)


def g_tag_before (dToken, dTags, sTag):
    if sTag not in dTags:
        return False
    if dToken["i"] > dTags[sTag][0]:
        return True
    return False


def g_tag_after (dToken, dTags, sTag):
    if sTag not in dTags:
        return False
    if dToken["i"] < dTags[sTag][1]:
        return True
    return False


#### Disambiguator

def select (dDA, nPos, sWord, sPattern, lDefault=None):
    if not sWord:
def g_select (dToken, sPattern, lDefault=None):
    "select morphologies for <dToken> according to <sPattern>, always return True"
    lMorph = dToken["lMorph"]  if "lMorph" in dToken  else _oSpellChecker.getMorph(dToken["sValue"])
    if not lMorph or len(lMorph) == 1:
        return True
    if nPos in dDA:
        if lDefault:
        return True
    if sWord not in _dAnalyses and not _storeMorphFromFSA(sWord):
        return True
            dToken["lMorph"] = lDefault
            #print("DA:", dToken["sValue"], dToken["lMorph"])
    if len(_dAnalyses[sWord]) == 1:
        return True
    lSelect = [ sMorph  for sMorph in _dAnalyses[sWord]  if re.search(sPattern, sMorph) ]
    lSelect = [ sMorph  for sMorph in lMorph  if re.search(sPattern, sMorph) ]
    if lSelect:
        if len(lSelect) != len(_dAnalyses[sWord]):
            dDA[nPos] = lSelect
        if len(lSelect) != len(lMorph):
            dToken["lMorph"] = lSelect
            #echo("= "+sWord+" "+str(dDA.get(nPos, "null")))
    elif lDefault:
        dDA[nPos] = lDefault
        dToken["lMorph"] = lDefault
        #echo("= "+sWord+" "+str(dDA.get(nPos, "null")))
    #print("DA:", dToken["sValue"], dToken["lMorph"])
    return True


def exclude (dDA, nPos, sWord, sPattern, lDefault=None):
    if not sWord:
def g_exclude (dToken, sPattern, lDefault=None):
    "select morphologies for <dToken> according to <sPattern>, always return True"
    lMorph = dToken["lMorph"]  if "lMorph" in dToken  else _oSpellChecker.getMorph(dToken["sValue"])
    if not lMorph or len(lMorph) == 1:
        if lDefault:
            dToken["lMorph"] = lDefault
            #print("DA:", dToken["sValue"], dToken["lMorph"])
        return True
    if nPos in dDA:
        return True
    if sWord not in _dAnalyses and not _storeMorphFromFSA(sWord):
        return True
    if len(_dAnalyses[sWord]) == 1:
        return True
    lSelect = [ sMorph  for sMorph in _dAnalyses[sWord]  if not re.search(sPattern, sMorph) ]
    lSelect = [ sMorph  for sMorph in lMorph  if not re.search(sPattern, sMorph) ]
    if lSelect:
        if len(lSelect) != len(_dAnalyses[sWord]):
            dDA[nPos] = lSelect
        if len(lSelect) != len(lMorph):
            dToken["lMorph"] = lSelect
            #echo("= "+sWord+" "+str(dDA.get(nPos, "null")))
    elif lDefault:
        dDA[nPos] = lDefault
        dToken["lMorph"] = lDefault
        #echo("= "+sWord+" "+str(dDA.get(nPos, "null")))
    #print("DA:", dToken["sValue"], dToken["lMorph"])
    return True


def define (dDA, nPos, lMorph):
    dDA[nPos] = lMorph
def g_define (dToken, lMorph):
    "set morphologies of <dToken>, always return True"
    dToken["lMorph"] = lMorph
    #echo("= "+str(nPos)+" "+str(dDA[nPos]))
    #print("DA:", dToken["sValue"], lMorph)
    return True



#### GRAMMAR CHECKER PLUGINS

${plugins}


#### CALLABLES FOR REGEX RULES (generated code)

${callables}


#### CALLABLES FOR GRAPH RULES (generated code)

${graph_callables}

"""
Grammar checker default options
"""

# generated code, do not edit

def getUI (sLang):
    "returns dictionary of UI labels"
    if sLang in _dOptLabel:
        return _dOptLabel[sLang]
    return _dOptLabel["fr"]


def getOptions (sContext="Python"):
    "returns dictionary of options"
    if sContext in dOpt:
        return dOpt[sContext]
    return dOpt["Python"]


lStructOpt = ${lStructOpt}

"""
Grammar checker regex rules
"""

# generated code, do not edit

lParagraphRules = ${paragraph_rules}

lSentenceRules = ${sentence_rules}

"""
Grammar checker graph rules
"""

# generated code, do not edit

dAllGraph = ${rules_graphs}

dRule = ${rules_actions}

#!python3

"""
Text tools
"""

import textwrap
from itertools import chain


def getParagraph (sText):
    "generator: returns paragraphs of text"

        return ""
    lGrammErrs = sorted(aGrammErrs, key=lambda d: d["nStart"])
    lSpellErrs = sorted(aSpellErrs, key=lambda d: d['nStart'])
    sText = ""
    nOffset = 0
    for sLine in wrap(sParagraph, nWidth): # textwrap.wrap(sParagraph, nWidth, drop_whitespace=False)
        sText += sLine + "\n"
        ln = len(sLine)
        nLineLen = len(sLine)
        sErrLine = ""
        nLenErrLine = 0
        nGrammErr = 0
        nSpellErr = 0
        for dErr in lGrammErrs:
            nStart = dErr["nStart"] - nOffset
            if nStart < ln:
            if nStart < nLineLen:
                nGrammErr += 1
                if nStart >= nLenErrLine:
                    sErrLine += " " * (nStart - nLenErrLine) + "^" * (dErr["nEnd"] - dErr["nStart"])
                    nLenErrLine = len(sErrLine)
            else:
                break
        for dErr in lSpellErrs:
            nStart = dErr['nStart'] - nOffset
            if nStart < ln:
            if nStart < nLineLen:
                nSpellErr += 1
                nEnd = dErr['nEnd'] - nOffset
                if nEnd > len(sErrLine):
                    sErrLine += " " * (nEnd - len(sErrLine))
                sErrLine = sErrLine[:nStart] + "°" * (nEnd - nStart) + sErrLine[nEnd:]
            else:
                break
        if sErrLine:
            sText += sErrLine + "\n"
        if nGrammErr:
            sText += getReadableErrors(lGrammErrs[:nGrammErr], nWidth)
            del lGrammErrs[0:nGrammErr]
        if nSpellErr:
            sText += getReadableErrors(lSpellErrs[:nSpellErr], nWidth, True)
            del lSpellErrs[0:nSpellErr]
        nOffset += ln
        nOffset += nLineLen
    return sText


def getReadableErrors (lErrs, nWidth, bSpell=False):
    "Returns lErrs errors as readable errors"
    sErrors = ""
    for dErr in lErrs:

    return sErrors


def getReadableError (dErr, bSpell=False):
    "Returns an error dErr as a readable error"
    try:
        if bSpell:
            s = u"* {nStart}:{nEnd}  # {sValue}:".format(**dErr)
            sText = u"* {nStart}:{nEnd}  # {sValue}:".format(**dErr)
        else:
            s = u"* {nStart}:{nEnd}  # {sLineId} / {sRuleId}:\n".format(**dErr)
            s += "  " + dErr.get("sMessage", "# error : message not found")
            sText = u"* {nStart}:{nEnd}  # {sLineId} / {sRuleId}:\n".format(**dErr)
            sText += "  " + dErr.get("sMessage", "# error : message not found")
        if dErr.get("aSuggestions", None):
            s += "\n  > Suggestions : " + " | ".join(dErr.get("aSuggestions", "# error : suggestions not found"))
            sText += "\n  > Suggestions : " + " | ".join(dErr.get("aSuggestions", "# error : suggestions not found"))
        if dErr.get("URL", None):
            s += "\n  > URL: " + dErr["URL"]
        return s
            sText += "\n  > URL: " + dErr["URL"]
        return sText
    except KeyError:
        return u"* Non-compliant error: {}".format(dErr)


def createParagraphWithLines (lLine):
    "Returns a text as merged lines and a set of data about lines (line_number_y, start_x, end_x)"
    sText = ""

# NOTES SUR LA LANGUE FRANÇAISE

## CE QUI ENTOURE UN VERBE

    PRONOMS (avant)
        COD         COI
        le / l’
        la / l’
        les
        en
        me / m’     me / m’
        te / t’     te / t’
        se / s’     lui
        nous        nous
        vous        nous
        se / s’     leur
                    y

    ADVERBE DE NÉGATION (avant)
        ne / n’

    SOMME
        [le|la|l’|les|en|me|m’|te|t’|se|s’|nous|vous|lui|leur|y]

    COMBINAISONS VALIDES
        ?[ne|n’]¿   [me|te|se]      [le|la|l’|les]
        ?[ne|n’]¿   [m’|t’|s’]      [le|la|l’|les|en|y]
        ?[ne|n’]¿   [le|la]         [lui|leur]
        ?[ne|n’]¿   [l’|les]        [lui|leur|en|y]
        ?[ne|n’]¿   [lui|leur]      en
        ?[ne|n’]¿   [nous|vous]     [le|la|l’|les|en|y]
        ne          [le|la|l’|les|me|m’|te|t’|se|s’|nous|vous|lui|leur]
        n’          [en|y]

    RÉSUMÉ & SIMPLIFICATION
        ?[ne|n’]¿   [le|la|l’|les|en|me|m’|te|t’|se|s’|nous|vous|lui|leur|y]
        ?[ne|n’]¿   [me|m’|te|t’|se|s’|nous|vous]   [le|la|l’|les|en|y]
        ?[ne|n’]¿   [le|la|l’|les]                  [lui|leur|en|y]
        ?[ne|n’]¿   [lui|leur]                      en

    ADVERBE DE NÉGATION (après)
        pas
        jamais
        point
        guère
        que / qu’
        rien

    PRONOMS À L’IMPÉRATIF
        APRÈS
            -moi
            -toi
            -lui
            -leur
            -nous
            -vous
            -le
            -la
            -les
            -en
            -y

        AVANT
            Uniquement les combinaisons avec l’adverbe de négation [ne|n’]

        return this._lVtyp[this._dVerb[sVerb][0]];
    },

    getSimil: function (sWord, sMorph, bSubst=false) {
        if (!sMorph.includes(":V")) {
            return new Set();
        }
        let sInfi = sMorph.slice(1, sMorph.indexOf(" "));
        let sInfi = sMorph.slice(1, sMorph.indexOf("/"));
        let aSugg = new Set();
        let tTags = this._getTags(sInfi);
        if (tTags) {
            if (!bSubst) {
                // we suggest conjugated forms
                if (sMorph.includes(":V1")) {
                    aSugg.add(sInfi);

    if (s2 == "vous") {
        return "vous";
    }
    if (s2 == "eux") {
        return "ils";
    }
    if (s2 == "elle" || s2 == "elles") {
        // We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
        if (cregex.mbNprMasNotFem(_dAnalyses.gl_get(s1, ""))) {
        if (cregex.mbNprMasNotFem(_oSpellChecker.getMorph(s1))) {
            return "ils";
        }
        // si épicène, indéterminable, mais OSEF, le féminin l’emporte
        return "elles";
    }
    return s1 + " et " + s2;
}

function apposition (sWord1, sWord2) {
    // returns true if nom + nom (no agreement required)
    // We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    return cregex.mbNomNotAdj(_dAnalyses.gl_get(sWord2, "")) && cregex.mbPpasNomNotAdj(_dAnalyses.gl_get(sWord1, ""));
    return cregex.mbNomNotAdj(_oSpellChecker.getMorph(sWord2)) && cregex.mbPpasNomNotAdj(_oSpellChecker.getMorph(sWord1));
}

function isAmbiguousNAV (sWord) {
    // words which are nom|adj and verb are ambiguous (except être and avoir)
    if (!_dAnalyses.has(sWord) && !_storeMorphFromFSA(sWord)) {
    let lMorph = _oSpellChecker.getMorph(sWord);
    if (lMorph.length === 0) {
        return false;
    }
    if (!cregex.mbNomAdj(_dAnalyses.gl_get(sWord, "")) || sWord == "est") {
    if (!cregex.mbNomAdj(lMorph) || sWord == "est") {
        return false;
    }
    if (cregex.mbVconj(_dAnalyses.gl_get(sWord, "")) && !cregex.mbMG(_dAnalyses.gl_get(sWord, ""))) {
    if (cregex.mbVconj(lMorph) && !cregex.mbMG(lMorph)) {
        return true;
    }
    return false;
}

function isAmbiguousAndWrong (sWord1, sWord2, sReqMorphNA, sReqMorphConj) {
    //// use it if sWord1 won’t be a verb; word2 is assumed to be true via isAmbiguousNAV
    // We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let a2 = _dAnalyses.gl_get(sWord2, null);
    if (!a2 || a2.length === 0) {
    let a2 = _oSpellChecker.getMorph(sWord2);
    if (a2.length === 0) {
        return false;
    }
    if (cregex.checkConjVerb(a2, sReqMorphConj)) {
        // verb word2 is ok
        return false;
    }
    let a1 = _dAnalyses.gl_get(sWord1, null);
    if (!a1 || a1.length === 0) {
    let a1 = _oSpellChecker.getMorph(sWord1);
    if (a1.length === 0) {
        return false;
    }
    if (cregex.checkAgreement(a1, a2) && (cregex.mbAdj(a2) || cregex.mbAdj(a1))) {
        return false;
    }
    return true;
}

function isVeryAmbiguousAndWrong (sWord1, sWord2, sReqMorphNA, sReqMorphConj, bLastHopeCond) {
    //// use it if sWord1 can be also a verb; word2 is assumed to be true via isAmbiguousNAV
    // We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let a2 = _dAnalyses.gl_get(sWord2, null);
    if (!a2 || a2.length === 0) {
    let a2 = _oSpellChecker.getMorph(sWord2);
    if (a2.length === 0) {
        return false;
    }
    if (cregex.checkConjVerb(a2, sReqMorphConj)) {
        // verb word2 is ok
        return false;
    }
    let a1 = _dAnalyses.gl_get(sWord1, null);
    if (!a1 || a1.length === 0) {
    let a1 = _oSpellChecker.getMorph(sWord1);
    if (a1.length === 0) {
        return false;
    }
    if (cregex.checkAgreement(a1, a2) && (cregex.mbAdj(a2) || cregex.mbAdjNb(a1))) {
        return false;
    }
    // now, we know there no agreement, and conjugation is also wrong
    if (cregex.isNomAdj(a1)) {
        return true;
    }
    //if cregex.isNomAdjVerb(a1): # considered true
    if (bLastHopeCond) {
        return true;
    }
    return false;
}

function checkAgreement (sWord1, sWord2) {
    // We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let a2 = _dAnalyses.gl_get(sWord2, null);
    if (!a2 || a2.length === 0) {
    let a2 = _oSpellChecker.getMorph(sWord2);
    if (a2.length === 0) {
        return true;
    }
    let a1 = _dAnalyses.gl_get(sWord1, null);
    if (!a1 || a1.length === 0) {
    let a1 = _oSpellChecker.getMorph(sWord1);
    if (a1.length === 0) {
        return true;
    }
    return cregex.checkAgreement(a1, a2);
}

function mbUnit (s) {
    if (/[µ\/⁰¹²³⁴⁵⁶⁷⁸⁹Ωℓ·]/.test(s)) {

    var phonet = require("resource://grammalecte/fr/phonet.js");
}


//// verbs

function suggVerb (sFlex, sWho, funcSugg2=null) {
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let aSugg = new Set();
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        let tTags = conj._getTags(sStem);
        if (tTags) {
            // we get the tense
            let aTense = new Set();
            for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
            for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
                let m;
                let zVerb = new RegExp (">"+sStem+" .*?(:(?:Y|I[pqsf]|S[pq]|K))", "g");
                let zVerb = new RegExp (">"+sStem+"/.*?(:(?:Y|I[pqsf]|S[pq]|K))", "g");
                while ((m = zVerb.exec(sMorph)) !== null) {
                    // stem must be used in regex to prevent confusion between different verbs (e.g. sauras has 2 stems: savoir and saurer)
                    if (m) {
                        if (m[1] === ":Y") {
                            aTense.add(":Ip");
                            aTense.add(":Iq");
                            aTense.add(":Is");

        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggVerbPpas (sFlex, sWhat=null) {
    let aSugg = new Set();
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        let tTags = conj._getTags(sStem);
        if (tTags) {
            if (!sWhat) {
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"));
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q2"));
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q3"));
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q4"));

        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggVerbTense (sFlex, sTense, sWho) {
    let aSugg = new Set();
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        if (conj.hasConj(sStem, sTense, sWho)) {
            aSugg.add(conj.getConj(sStem, sTense, sWho));
        }
    }
    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggVerbImpe (sFlex) {
    let aSugg = new Set();
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        let tTags = conj._getTags(sStem);
        if (tTags) {
            if (conj._hasConjWithTags(tTags, ":E", ":2s")) {
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":E", ":2s"));
            }
            if (conj._hasConjWithTags(tTags, ":E", ":1p")) {
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":E", ":1p"));
            }
            if (conj._hasConjWithTags(tTags, ":E", ":2p")) {
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":E", ":2p"));
            }
        }
    }
    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggVerbInfi (sFlex) {
    return stem(sFlex).filter(sStem => conj.isVerb(sStem)).join("|");
    return _oSpellChecker.getLemma(sFlex).filter(sStem => conj.isVerb(sStem)).join("|");
}


const _dQuiEst = new Map ([
    ["je", ":1s"], ["j’", ":1s"], ["j’en", ":1s"], ["j’y", ":1s"],
    ["tu", ":2s"], ["il", ":3s"], ["on", ":3s"], ["elle", ":3s"],
    ["nous", ":1p"], ["vous", ":2p"], ["ils", ":3p"], ["elles", ":3p"]

    if (!sWho) {
        if (sSuj[0].gl_isLowerCase()) { // pas un pronom, ni un nom propre
            return "";
        }
        sWho = ":3s";
    }
    let aSugg = new Set();
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        let tTags = conj._getTags(sStem);
        if (tTags) {
            for (let sTense of lMode) {
                if (conj._hasConjWithTags(tTags, sTense, sWho)) {
                    aSugg.add(conj._getConjWithTags(sStem, tTags, sTense, sWho));
                }
            }
        }
    }
    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }
    return "";
}

//// Nouns and adjectives

function suggPlur (sFlex, sWordToAgree=null) {
    // returns plural forms assuming sFlex is singular
    if (sWordToAgree) {
        if (!_dAnalyses.has(sWordToAgree) && !_storeMorphFromFSA(sWordToAgree)) {
        let lMorph = _oSpellChecker.getMorph(sWordToAgree);
        if (lMorph.length === 0) {
            return "";
        }
        let sGender = cregex.getGender(_dAnalyses.gl_get(sWordToAgree, []));
        let sGender = cregex.getGender(lMorph);
        if (sGender == ":m") {
            return suggMasPlur(sFlex);
        } else if (sGender == ":f") {
            return suggFemPlur(sFlex);
        }
    }
    let aSugg = new Set();

        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggMasSing (sFlex, bSuggSimil=false) {
    // returns masculine singular forms
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let aSugg = new Set();
    for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
    for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
        if (!sMorph.includes(":V")) {
            // not a verb
            if (sMorph.includes(":m") || sMorph.includes(":e")) {
                aSugg.add(suggSing(sFlex));
            } else {
                let sStem = cregex.getLemmaOfMorph(sMorph);
                if (mfsp.isFemForm(sStem)) {

        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggMasPlur (sFlex, bSuggSimil=false) {
    // returns masculine plural forms
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let aSugg = new Set();
    for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
    for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
        if (!sMorph.includes(":V")) {
            // not a verb
            if (sMorph.includes(":m") || sMorph.includes(":e")) {
                aSugg.add(suggPlur(sFlex));
            } else {
                let sStem = cregex.getLemmaOfMorph(sMorph);
                if (mfsp.isFemForm(sStem)) {

    }
    return "";
}


function suggFemSing (sFlex, bSuggSimil=false) {
    // returns feminine singular forms
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let aSugg = new Set();
    for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
    for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
        if (!sMorph.includes(":V")) {
            // not a verb
            if (sMorph.includes(":f") || sMorph.includes(":e")) {
                aSugg.add(suggSing(sFlex));
            } else {
                let sStem = cregex.getLemmaOfMorph(sMorph);
                if (mfsp.isFemForm(sStem)) {

        return Array.from(aSugg).join("|");
    }
    return "";
}

function suggFemPlur (sFlex, bSuggSimil=false) {
    // returns feminine plural forms
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let aSugg = new Set();
    for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
    for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
        if (!sMorph.includes(":V")) {
            // not a verb
            if (sMorph.includes(":f") || sMorph.includes(":e")) {
                aSugg.add(suggPlur(sFlex));
            } else {
                let sStem = cregex.getLemmaOfMorph(sMorph);
                if (mfsp.isFemForm(sStem)) {

    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }
    return "";
}

function hasFemForm (sFlex) {
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        if (mfsp.isFemForm(sStem) || conj.hasConj(sStem, ":PQ", ":Q3")) {
            return true;
        }
    }
    if (phonet.hasSimil(sFlex, ":f")) {
        return true;
    }
    return false;
}

function hasMasForm (sFlex) {
    for (let sStem of stem(sFlex)) {
    for (let sStem of _oSpellChecker.getLemma(sFlex)) {
        if (mfsp.isFemForm(sStem) || conj.hasConj(sStem, ":PQ", ":Q1")) {
            // what has a feminine form also has a masculine form
            return true;
        }
    }
    if (phonet.hasSimil(sFlex, ":m")) {
        return true;
    }
    return false;
}

function switchGender (sFlex, bPlur=null) {
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    let aSugg = new Set();
    if (bPlur === null) {
        for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
        for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
            if (sMorph.includes(":f")) {
                if (sMorph.includes(":s")) {
                    aSugg.add(suggMasSing(sFlex));
                } else if (sMorph.includes(":p")) {
                    aSugg.add(suggMasPlur(sFlex));
                }
            } else if (sMorph.includes(":m")) {
                if (sMorph.includes(":s")) {
                    aSugg.add(suggFemSing(sFlex));
                } else if (sMorph.includes(":p")) {
                    aSugg.add(suggFemPlur(sFlex));
                } else {
                    aSugg.add(suggFemSing(sFlex));
                    aSugg.add(suggFemPlur(sFlex));
                }
            }
        }
    } else if (bPlur) {
        for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
        for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
            if (sMorph.includes(":f")) {
                aSugg.add(suggMasPlur(sFlex));
            } else if (sMorph.includes(":m")) {
                aSugg.add(suggFemPlur(sFlex));
            }
        }
    } else {
        for (let sMorph of _dAnalyses.gl_get(sFlex, [])) {
        for (let sMorph of _oSpellChecker.getMorph(sFlex)) {
            if (sMorph.includes(":f")) {
                aSugg.add(suggMasSing(sFlex));
            } else if (sMorph.includes(":m")) {
                aSugg.add(suggFemSing(sFlex));
            }
        }
    }
    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }
    return "";
}

function switchPlural (sFlex) {
    let aSugg = new Set();
    for (let sMorph of _dAnalyses.gl_get(sFlex, [])) { // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    for (let sMorph of _oSpellChecker.getMorph(sFlex)) { 
        if (sMorph.includes(":s")) {
            aSugg.add(suggPlur(sFlex));
        } else if (sMorph.includes(":p")) {
            aSugg.add(suggSing(sFlex));
        }
    }
    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }
    return "";
}

function hasSimil (sWord, sPattern=null) {
    return phonet.hasSimil(sWord, sPattern);
}

function suggSimil (sWord, sPattern=null, bSubst=false) {
    // return list of words phonetically similar to sWord and whom POS is matching sPattern
    let aSugg = phonet.selectSimil(sWord, sPattern);
    for (let sMorph of _dAnalyses.gl_get(sWord, [])) {
    for (let sMorph of _oSpellChecker.getMorph(sWord)) {
        for (let e of conj.getSimil(sWord, sMorph, bSubst)) {
            aSugg.add(e);
        }
    }
    if (aSugg.size > 0) {
        return Array.from(aSugg).join("|");
    }

    if (sWord[0] == "h" || sWord[0] == "H") {
        return "ce|cet";
    }
    return "ce";
}

function suggLesLa (sWord) {
    // we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    if (_dAnalyses.gl_get(sWord, []).some(s  =>  s.includes(":p"))) {
    if (_oSpellChecker.getMorph(sWord).some(s  =>  s.includes(":p"))) {
        return "les|la";
    }
    return "la";
}

function formatNumber (s) {
    let nLen = s.length;

${gctestsJS}
${regex_gctestsJS}

"""
# Grammalecte - Conjugueur
Grammalecte - Conjugueur
"""

# License: GPL 3

import re
import traceback

from .conj_data import lVtyp as _lVtyp
from .conj_data import lTags as _lTags

_dGroup = { "0": "auxiliaire", "1": "1ᵉʳ groupe", "2": "2ᵉ groupe", "3": "3ᵉ groupe" }

_dTenseIdx = { ":PQ": 0, ":Ip": 1, ":Iq": 2, ":Is": 3, ":If": 4, ":K": 5, ":Sp": 6, ":Sq": 7, ":E": 8 }



def isVerb (sVerb):
    "return True if it’s a existing verb"
    return sVerb in _dVerb


def getConj (sVerb, sTense, sWho):
    "returns conjugation (can be an empty string)"
    if sVerb not in _dVerb:
        return None

    "returns raw informations about sVerb"
    if sVerb not in _dVerb:
        return None
    return _lVtyp[_dVerb[sVerb][0]]


def getSimil (sWord, sMorph, bSubst=False):
    "returns a set of verbal forms similar to <sWord>, according to <sMorph>"
    if ":V" not in sMorph:
        return set()
    sInfi = sMorph[1:sMorph.find(" ")]
    sInfi = sMorph[1:sMorph.find("/")]
    aSugg = set()
    tTags = _getTags(sInfi)
    if tTags:
        if not bSubst:
            # we suggest conjugated forms
            if ":V1" in sMorph:
                aSugg.add(sInfi)

            # if there is only one past participle (epi inv), unreliable.
            if len(aSugg) == 1:
                aSugg.clear()
    return aSugg


def getConjSimilInfiV1 (sInfi):
    "returns verbal forms phonetically similar to infinitive form (for verb in group 1)"
    if sInfi not in _dVerb:
        return set()
    aSugg = set()
    tTags = _getTags(sInfi)
    if tTags:
        aSugg.add(_getConjWithTags(sInfi, tTags, ":Iq", ":2s"))
        aSugg.add(_getConjWithTags(sInfi, tTags, ":Iq", ":3s"))

    "returns sWord modified by sSfx"
    if not sSfx:
        return ""
    if sSfx == "0":
        return sWord
    try:
        return sWord[:-(ord(sSfx[0])-48)] + sSfx[1:]  if sSfx[0] != '0'  else  sWord + sSfx[1:]  # 48 is the ASCII code for "0"
    except:
    except (IndexError, TypeError):
        return "## erreur, code : " + str(sSfx) + " ##"
        



class Verb ():
    "Verb and its conjugation"

    def __init__ (self, sVerb, sVerbPattern=""):
        # conjugate a unknown verb with rules from sVerbPattern
        if not isinstance(sVerb, str):
            raise TypeError("sVerb should be a string")
        if not sVerb:
            raise ValueError("Empty string.")

        if sVerbPattern == "":
            sVerbPattern = sVerb

        self.sVerb = sVerb
        self.sVerbAux = ""
        self._sRawInfo = getVtyp(sVerbPattern)
        self.sInfo = self._readableInfo()
        self.bProWithEn = (self._sRawInfo[5] == "e")
        self._tTags = _getTags(sVerbPattern)
        if not self._tTags:
            raise ValueError("Unknown verb.")
        self._tTagsAux = _getTags(self.sVerbAux)
        self.cGroup = self._sRawInfo[0];
        self.cGroup = self._sRawInfo[0]
        self.dConj = {
            ":Y": {
                "label": "Infinitif",
                ":": sVerb,
            },
            ":P": {
                "label": "Participe présent",

                sInfo = "# erreur - code : " + self._sRawInfo
            return sGroup + " · " + sInfo
        except:
            traceback.print_exc()
            return "# erreur"

    def infinitif (self, bPro, bNeg, bTpsCo, bInt, bFem):
        "returns string (conjugaison à l’infinitif)"
        try:
            if bTpsCo:
                sInfi = self.sVerbAux  if not bPro  else  "être"
            else:
                sInfi = self.sVerb
            if bPro:
                if self.bProWithEn:

                sInfi += " … ?"
            return sInfi
        except:
            traceback.print_exc()
            return "# erreur"

    def participePasse (self, sWho):
        "returns past participle according to <sWho>"
        try:
            return self.dConj[":Q"][sWho]
        except:
            traceback.print_exc()
            return "# erreur"

    def participePresent (self, bPro, bNeg, bTpsCo, bInt, bFem):
        "returns string (conjugaison du participe présent)"
        try:
            if not self.dConj[":P"][":"]:
                return ""
            if bTpsCo:
                sPartPre = _getConjWithTags(self.sVerbAux, self._tTagsAux, ":PQ", ":P")  if not bPro  else  getConj("être", ":PQ", ":P")
            else:
                sPartPre = self.dConj[":P"][":"]

                sPartPre += " … ?"
            return sPartPre
        except:
            traceback.print_exc()
            return "# erreur"

    def conjugue (self, sTemps, sWho, bPro, bNeg, bTpsCo, bInt, bFem):
        "returns string (conjugue le verbe au temps <sTemps> pour <sWho>) "
        try:
            if not self.dConj[sTemps][sWho]:
                return ""
            if not bTpsCo and bInt and sWho == ":1s" and self.dConj[sTemps].get(":1ś", False):
                sWho = ":1ś"
            if bTpsCo:
                sConj = _getConjWithTags(self.sVerbAux, self._tTagsAux, sTemps, sWho)  if not bPro  else  getConj("être", sTemps, sWho)

                else:
                    sConj = _dProObjEl[sWho] + "en " + sConj
            if bNeg:
                sConj = "n’" + sConj  if bEli and not bPro  else  "ne " + sConj
            if bInt:
                if sWho == ":3s" and not _zNeedTeuph.search(sConj):
                    sConj += "-t"
                sConj += "-" + self._getPronom(sWho, bFem)
                sConj += "-" + self._getPronomSujet(sWho, bFem)
            else:
                if sWho == ":1s" and bEli and not bNeg and not bPro:
                    sConj = "j’" + sConj
                else:
                    sConj = self._getPronom(sWho, bFem) + " " + sConj
                    sConj = self._getPronomSujet(sWho, bFem) + " " + sConj
            if bNeg:
                sConj += " pas"
            if bTpsCo:
                sConj += " " + self._seekPpas(bPro, bFem, sWho.endswith("p") or self._sRawInfo[5] == "r")
            if bInt:
                sConj += " … ?"
            return sConj
        except:
            traceback.print_exc()
            return "# erreur"

    def _getPronom (self, sWho, bFem):
    def _getPronomSujet (self, sWho, bFem):
        try:
            if sWho == ":3s":
                if self._sRawInfo[5] == "r":
                    return "on"
                elif bFem:
                    return "elle"
            elif sWho == ":3p" and bFem:
                return "elles"
            return _dProSuj[sWho]
        except:
            traceback.print_exc()
            return "# erreur"

    def imperatif (self, sWho, bPro, bNeg, bTpsCo, bFem):
        "returns string (conjugaison à l’impératif)"
        try:
            if not self.dConj[":E"][sWho]:
                return ""
            if bTpsCo:
                sImpe = _getConjWithTags(self.sVerbAux, self._tTagsAux, ":E", sWho)  if not bPro  else  getConj(u"être", ":E", sWho)
            else:
                sImpe = self.dConj[":E"][sWho]

"""
# Conjugation generator
# beta stage, unfinished, the root for a new way to generate flexions…
Conjugation generator
beta stage, unfinished, the root for a new way to generate flexions…
"""

import re


def conjugate (sVerb, sVerbTag="i_____a", bVarPpas=True):
    "conjugate <sVerb> and returns a list of tuples (conjugation form, tags)"
    lConj = []
    cGroup = getVerbGroupChar(sVerb)
    for nCut, sAdd, sFlexTags, sPattern in getConjRules(sVerb, bVarPpas):
        if not sPattern or re.search(sPattern, sVerb):
            sFlexion = sVerb[0:-nCut] + sAdd  if nCut  else sVerb + sAdd
            lConj.append((sFlexion, ":V" + cGroup + "_" + sVerbTag + sFlexTags))
    return lConj


def getVerbGroupChar (sVerb, ):
def getVerbGroupChar (sVerb):
    "returns the group number of <sVerb> guessing on its ending"
    sVerb = sVerb.lower()
    if sVerb.endswith("er"):
        return "1"
    if sVerb.endswith("ir"):
        return "2"
    if sVerb == "être" or sVerb == "avoir":
        return "0"
    if sVerb.endswith("re"):
        return "3"
    return "4"


def getConjRules (sVerb, bVarPpas=True, nGroup=2):
    "returns a list of lists to conjugate a verb, guessing on its ending"
    if sVerb.endswith("er"):
        # premier groupe, conjugaison en fonction de la terminaison du lemme
        # 5 lettres
        if sVerb[-5:] in oConj["V1"]:
            lConj = list(oConj["V1"][sVerb[-5:]])
        # 4 lettres
        elif sVerb[-4:] in oConj["V1"]:

        [2,     "isses",        ":Sp:Sq:2s/*",      False],
        [2,     "isse",         ":Sp:3s/*",         False],
        [2,     "ît",           ":Sq:3s/*",         False],
        [2,     "is",           ":E:2s/*",          False],
        [2,     "issons",       ":E:1p/*",          False],
        [2,     "issez",        ":E:2p/*",          False]
    ],
    

    # premier groupe (bien plus irrégulier que prétendu)
    "V1": {
        # a
        # verbes en -er, -ger, -yer, -cer
        "er": [
            [2,      "er",        ":Y/*",               False],
            [2,      "ant",       ":P/*",               False],

"""
# Grammalecte - Compiled regular expressions
Grammalecte - Compiled regular expressions
"""

import re

#### Lemme
Lemma = re.compile("^>(\w[\w-]*)")
Lemma = re.compile(r"^>(\w[\w-]*)")

#### Analyses
Gender = re.compile(":[mfe]")
Number = re.compile(":[spi]")

#### Nom et adjectif
NA = re.compile(":[NA]")

NPf = re.compile(":(?:M[12P]|T):f")
NPe = re.compile(":(?:M[12P]|T):e")


#### FONCTIONS

def getLemmaOfMorph (s):
    "return lemma in morphology <s>"
    return Lemma.search(s).group(1)

def checkAgreement (l1, l2):
    "returns True if agreement in gender and number is possible between morphologies <l1> and <l2>"
    # check number agreement
    if not mbInv(l1) and not mbInv(l2):
        if mbSg(l1) and not mbSg(l2):
            return False
        if mbPl(l1) and not mbPl(l2):
            return False
    # check gender agreement
    if mbEpi(l1) or mbEpi(l2):
        return True
    if mbMas(l1) and not mbMas(l2):
        return False
    if mbFem(l1) and not mbFem(l2):
        return False
    return True

def checkConjVerb (lMorph, sReqConj):
    "returns True if <sReqConj> in <lMorph>"
    return any(sReqConj in s  for s in lMorph)

def getGender (lMorph):
    "returns gender of word (':m', ':f', ':e' or empty string)."
    sGender = ""
    for sMorph in lMorph:
        m = Gender.search(sMorph)
        if m:
            if not sGender:
                sGender = m.group(0)
            elif sGender != m.group(0):
                return ":e"
    return sGender

def getNumber (lMorph):
    "returns number of word (':s', ':p', ':i' or empty string)."
    sNumber = ""
    for sMorph in lMorph:
        m = Number.search(sWord)
        m = Number.search(sMorph)
        if m:
            if not sNumber:
                sNumber = m.group(0)
            elif sNumber != m.group(0):
                return ":i"
    return sNumber

# NOTE :  isWhat (lMorph)    returns True   if lMorph contains nothing else than What
#         mbWhat (lMorph)    returns True   if lMorph contains What at least once

## isXXX = it’s certain

def isNom (lMorph):
    "returns True if all morphologies are “nom”"
    return all(":N" in s  for s in lMorph)

def isNomNotAdj (lMorph):
    "returns True if all morphologies are “nom”, but not “adjectif”"
    return all(NnotA.search(s)  for s in lMorph)

def isAdj (lMorph):
    "returns True if all morphologies are “adjectif”"
    return all(":A" in s  for s in lMorph)

def isNomAdj (lMorph):
    "returns True if all morphologies are “nom” or “adjectif”"
    return all(NA.search(s)  for s in lMorph)

def isNomVconj (lMorph):
    "returns True if all morphologies are “nom” or “verbe conjugué”"
    return all(NVconj.search(s)  for s in lMorph)

def isInv (lMorph):
    "returns True if all morphologies are “invariable”"
    return all(":i" in s  for s in lMorph)

def isSg (lMorph):
    "returns True if all morphologies are “singulier”"
    return all(":s" in s  for s in lMorph)

def isPl (lMorph):
    "returns True if all morphologies are “pluriel”"
    return all(":p" in s  for s in lMorph)

def isEpi (lMorph):
    "returns True if all morphologies are “épicène”"
    return all(":e" in s  for s in lMorph)

def isMas (lMorph):
    "returns True if all morphologies are “masculin”"
    return all(":m" in s  for s in lMorph)

def isFem (lMorph):
    "returns True if all morphologies are “féminin”"
    return all(":f" in s  for s in lMorph)


## mbXXX = MAYBE XXX

def mbNom (lMorph):
    "returns True if one morphology is “nom”"
    return any(":N" in s  for s in lMorph)

def mbAdj (lMorph):
    "returns True if one morphology is “adjectif”"
    return any(":A" in s  for s in lMorph)

def mbAdjNb (lMorph):
    "returns True if one morphology is “adjectif” or “nombre”"
    return any(AD.search(s)  for s in lMorph)

def mbNomAdj (lMorph):
    "returns True if one morphology is “nom” or “adjectif”"
    return any(NA.search(s)  for s in lMorph)

def mbNomNotAdj (lMorph):
    "returns True if one morphology is “nom”, but not “adjectif”"
    b = False
    bResult = False
    for s in lMorph:
        if ":A" in s:
            return False
        if ":N" in s:
            b = True
    return b
            bResult = True
    return bResult

def mbPpasNomNotAdj (lMorph):
    "returns True if one morphology is “nom” or “participe passé”, but not “adjectif”"
    return any(PNnotA.search(s)  for s in lMorph)

def mbVconj (lMorph):
    "returns True if one morphology is “nom” or “verbe conjugué”"
    return any(Vconj.search(s)  for s in lMorph)

def mbVconj123 (lMorph):
    "returns True if one morphology is “nom” or “verbe conjugué” (but not “avoir” or “être”)"
    return any(Vconj123.search(s)  for s in lMorph)

def mbMG (lMorph):
    "returns True if one morphology is “mot grammatical”"
    return any(":G" in s  for s in lMorph)

def mbInv (lMorph):
    "returns True if one morphology is “invariable”"
    return any(":i" in s  for s in lMorph)

def mbSg (lMorph):
    "returns True if one morphology is “singulier”"
    return any(":s" in s  for s in lMorph)

def mbPl (lMorph):
    "returns True if one morphology is “pluriel”"
    return any(":p" in s  for s in lMorph)

def mbEpi (lMorph):
    "returns True if one morphology is “épicène”"
    return any(":e" in s  for s in lMorph)

def mbMas (lMorph):
    "returns True if one morphology is “masculin”"
    return any(":m" in s  for s in lMorph)

def mbFem (lMorph):
    "returns True if one morphology is “féminin”"
    return any(":f" in s  for s in lMorph)

def mbNpr (lMorph):
    "returns True if one morphology is “nom propre” or “titre de civilité”"
    return any(NP.search(s)  for s in lMorph)

def mbNprMasNotFem (lMorph):
    "returns True if one morphology is “nom propre masculin” but not “féminin”"
    if any(NPf.search(s)  for s in lMorph):
        return False
    return any(NPm.search(s)  for s in lMorph)

#### GRAMMAR CHECKING ENGINE PLUGIN: Parsing functions for French language

from . import cregex as cr


def rewriteSubject (s1, s2):
    # s1 is supposed to be prn/patr/npr (M[12P])
    "rewrite complex subject: <s1> a prn/patr/npr (M[12P]) followed by “et” and <s2>"
    if s2 == "lui":
        return "ils"
    if s2 == "moi":
        return "nous"
    if s2 == "toi":
        return "vous"
    if s2 == "nous":
        return "nous"
    if s2 == "vous":
        return "vous"
    if s2 == "eux":
        return "ils"
    if s2 == "elle" or s2 == "elles":
        # We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
        if cr.mbNprMasNotFem(_dAnalyses.get(s1, False)):
        if cr.mbNprMasNotFem(_oSpellChecker.getMorph(s1)):
            return "ils"
        # si épicène, indéterminable, mais OSEF, le féminin l’emporte
        return "elles"
    return s1 + " et " + s2


def apposition (sWord1, sWord2):
    "returns True if nom + nom (no agreement required)"
    # We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    return cr.mbNomNotAdj(_dAnalyses.get(sWord2, False)) and cr.mbPpasNomNotAdj(_dAnalyses.get(sWord1, False))
    return cr.mbNomNotAdj(_oSpellChecker.getMorph(sWord2)) and cr.mbPpasNomNotAdj(_oSpellChecker.getMorph(sWord1))


def isAmbiguousNAV (sWord):
    "words which are nom|adj and verb are ambiguous (except être and avoir)"
    if sWord not in _dAnalyses and not _storeMorphFromFSA(sWord):
    lMorph = _oSpellChecker.getMorph(sWord)
        return False
    if not cr.mbNomAdj(_dAnalyses[sWord]) or sWord == "est":
    if not cr.mbNomAdj(lMorph) or sWord == "est":
        return False
    if cr.mbVconj(_dAnalyses[sWord]) and not cr.mbMG(_dAnalyses[sWord]):
    if cr.mbVconj(lMorph) and not cr.mbMG(lMorph):
        return True
    return False


def isAmbiguousAndWrong (sWord1, sWord2, sReqMorphNA, sReqMorphConj):
    "use it if sWord1 won’t be a verb; word2 is assumed to be True via isAmbiguousNAV"
    "use it if <sWord1> won’t be a verb; <sWord2> is assumed to be True via isAmbiguousNAV"
    # We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    a2 = _dAnalyses.get(sWord2, None)
    a2 = _oSpellChecker.getMorph(sWord2)
    if not a2:
        return False
    if cr.checkConjVerb(a2, sReqMorphConj):
        # verb word2 is ok
        return False
    a1 = _dAnalyses.get(sWord1, None)
    a1 = _oSpellChecker.getMorph(sWord1)
    if not a1:
        return False
    if cr.checkAgreement(a1, a2) and (cr.mbAdj(a2) or cr.mbAdj(a1)):
        return False
    return True


def isVeryAmbiguousAndWrong (sWord1, sWord2, sReqMorphNA, sReqMorphConj, bLastHopeCond):
    "use it if sWord1 can be also a verb; word2 is assumed to be True via isAmbiguousNAV"
    "use it if <sWord1> can be also a verb; <sWord2> is assumed to be True via isAmbiguousNAV"
    # We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    a2 = _dAnalyses.get(sWord2, None)
    a2 = _oSpellChecker.getMorph(sWord2)
    if not a2:
        return False
    if cr.checkConjVerb(a2, sReqMorphConj):
        # verb word2 is ok
        return False
    a1 = _dAnalyses.get(sWord1, None)
    a1 = _oSpellChecker.getMorph(sWord1)
    if not a1:
        return False
    if cr.checkAgreement(a1, a2) and (cr.mbAdj(a2) or cr.mbAdjNb(a1)):
        return False
    # now, we know there no agreement, and conjugation is also wrong
    if cr.isNomAdj(a1):
        return True
    #if cr.isNomAdjVerb(a1): # considered True
    if bLastHopeCond:
        return True
    return False


def checkAgreement (sWord1, sWord2):
    # We don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    a2 = _dAnalyses.get(sWord2, None)
    "check agreement between <sWord1> and <sWord1>"
    a2 = _oSpellChecker.getMorph(sWord2)
    if not a2:
        return True
    a1 = _dAnalyses.get(sWord1, None)
    a1 = _oSpellChecker.getMorph(sWord1)
    if not a1:
        return True
    return cr.checkAgreement(a1, a2)


_zUnitSpecial = re.compile("[µ/⁰¹²³⁴⁵⁶⁷⁸⁹Ωℓ·]")
_zUnitNumbers = re.compile("[0-9]")

def mbUnit (s):
    "returns True it can be a measurement unit"
    if _zUnitSpecial.search(s):
        return True
    if 1 < len(s) < 16 and s[0:1].islower() and (not s[1:].islower() or _zUnitNumbers.search(s)):
        return True
    return False


#### Syntagmes

_zEndOfNG1 = re.compile(" *$| +(?:, +|)(?:n(?:’|e |o(?:u?s|tre) )|l(?:’|e(?:urs?|s|) |a )|j(?:’|e )|m(?:’|es? |a |on )|t(?:’|es? |a |u )|s(?:’|es? |a )|c(?:’|e(?:t|tte|s|) )|ç(?:a |’)|ils? |vo(?:u?s|tre) )")
_zEndOfNG2 = re.compile(r" +(\w[\w-]+)")
_zEndOfNG3 = re.compile(r" *, +(\w[\w-]+)")

def isEndOfNG (dDA, s, iOffset):
    "returns True if next word doesn’t belong to a noun group"
    if _zEndOfNG1.match(s):
        return True
    m = _zEndOfNG2.match(s)
    if m and morphex(dDA, (iOffset+m.start(1), m.group(1)), ":[VR]", ":[NAQP]"):
        return True
    m = _zEndOfNG3.match(s)
    if m and not morph(dDA, (iOffset+m.start(1), m.group(1)), ":[NA]", False):
        return True
    return False


_zNextIsNotCOD1 = re.compile(" *,")
_zNextIsNotCOD2 = re.compile(" +(?:[mtsnj](e +|’)|[nv]ous |tu |ils? |elles? )")
_zNextIsNotCOD3 = re.compile(r" +([a-zéèî][\w-]+)")

def isNextNotCOD (dDA, s, iOffset):
    "returns True if next word is not a COD"
    if _zNextIsNotCOD1.match(s) or _zNextIsNotCOD2.match(s):
        return True
    m = _zNextIsNotCOD3.match(s)
    if m and morphex(dDA, (iOffset+m.start(1), m.group(1)), ":[123][sp]", ":[DM]"):
        return True
    return False


_zNextIsVerb1 = re.compile(" +[nmts](?:e |’)")
_zNextIsVerb2 = re.compile(r" +(\w[\w-]+)")

def isNextVerb (dDA, s, iOffset):
    "returns True if next word is a verb"
    if _zNextIsVerb1.match(s):
        return True
    m = _zNextIsVerb2.match(s)
    if m and morph(dDA, (iOffset+m.start(1), m.group(1)), ":[123][sp]", False):
        return True
    return False


#### Exceptions

aREGULARPLURAL = frozenset(["abricot", "amarante", "aubergine", "acajou", "anthracite", "brique", "caca", "café", \
                            "carotte", "cerise", "chataigne", "corail", "citron", "crème", "grave", "groseille", \
                            "jonquille", "marron", "olive", "pervenche", "prune", "sable"])
aSHOULDBEVERB = frozenset(["aller", "manger"]) 
aSHOULDBEVERB = frozenset(["aller", "manger"])

#### GRAMMAR CHECKING ENGINE PLUGIN: Suggestion mechanisms

from . import conj
from . import mfsp
from . import phonet


## Verbs

def suggVerb (sFlex, sWho, funcSugg2=None):
    "change <sFlex> conjugation according to <sWho>"
    aSugg = set()
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        tTags = conj._getTags(sStem)
        if tTags:
            # we get the tense
            aTense = set()
            for sMorph in _dAnalyses.get(sFlex, []): # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
                for m in re.finditer(">"+sStem+" .*?(:(?:Y|I[pqsf]|S[pq]|K|P))", sMorph):
            for sMorph in _oSpellChecker.getMorph(sFlex):
                for m in re.finditer(">"+sStem+"/.*?(:(?:Y|I[pqsf]|S[pq]|K|P))", sMorph):
                    # stem must be used in regex to prevent confusion between different verbs (e.g. sauras has 2 stems: savoir and saurer)
                    if m:
                        if m.group(1) == ":Y":
                            aTense.add(":Ip")
                            aTense.add(":Iq")
                            aTense.add(":Is")
                        elif m.group(1) == ":P":

        if aSugg2:
            aSugg.add(aSugg2)
    if aSugg:
        return "|".join(aSugg)
    return ""


def suggVerbPpas (sFlex, sWhat=None):
def suggVerbPpas (sFlex, sPattern=None):
    "suggest past participles for <sFlex>"
    aSugg = set()
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        tTags = conj._getTags(sStem)
        if tTags:
            if not sWhat:
            if not sPattern:
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q2"))
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q3"))
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q4"))
                aSugg.discard("")
            elif sWhat == ":m:s":
            elif sPattern == ":m:s":
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
            elif sWhat == ":m:p":
            elif sPattern == ":m:p":
                if conj._hasConjWithTags(tTags, ":PQ", ":Q2"):
                    aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q2"))
                else:
                    aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
            elif sWhat == ":f:s":
            elif sPattern == ":f:s":
                if conj._hasConjWithTags(tTags, ":PQ", ":Q3"):
                    aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q3"))
                else:
                    aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
            elif sWhat == ":f:p":
            elif sPattern == ":f:p":
                if conj._hasConjWithTags(tTags, ":PQ", ":Q4"):
                    aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q4"))
                else:
                    aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
            elif sWhat == ":s":
            elif sPattern == ":s":
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q3"))
                aSugg.discard("")
            elif sWhat == ":p":
            elif sPattern == ":p":
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q2"))
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q4"))
                aSugg.discard("")
            else:
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":PQ", ":Q1"))
    if aSugg:
        return "|".join(aSugg)
    return ""


def suggVerbTense (sFlex, sTense, sWho):
    "change <sFlex> to a verb according to <sTense> and <sWho>"
    aSugg = set()
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        if conj.hasConj(sStem, sTense, sWho):
            aSugg.add(conj.getConj(sStem, sTense, sWho))
    if aSugg:
        return "|".join(aSugg)
    return ""


def suggVerbImpe (sFlex):
    "change <sFlex> to a verb at imperative form"
    aSugg = set()
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        tTags = conj._getTags(sStem)
        if tTags:
            if conj._hasConjWithTags(tTags, ":E", ":2s"):
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":E", ":2s"))
            if conj._hasConjWithTags(tTags, ":E", ":1p"):
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":E", ":1p"))
            if conj._hasConjWithTags(tTags, ":E", ":2p"):
                aSugg.add(conj._getConjWithTags(sStem, tTags, ":E", ":2p"))
    if aSugg:
        return "|".join(aSugg)
    return ""


def suggVerbInfi (sFlex):
    "returns infinitive forms of <sFlex>"
    return "|".join([ sStem  for sStem in stem(sFlex)  if conj.isVerb(sStem) ])
    return "|".join([ sStem  for sStem in _oSpellChecker.getLemma(sFlex)  if conj.isVerb(sStem) ])


_dQuiEst = { "je": ":1s", "j’": ":1s", "j’en": ":1s", "j’y": ":1s", \
             "tu": ":2s", "il": ":3s", "on": ":3s", "elle": ":3s", "nous": ":1p", "vous": ":2p", "ils": ":3p", "elles": ":3p" }
_lIndicatif = [":Ip", ":Iq", ":Is", ":If"]
_lSubjonctif = [":Sp", ":Sq"]

def suggVerbMode (sFlex, cMode, sSuj):
    "returns other conjugations of <sFlex> acconding to <cMode> and <sSuj>"
    if cMode == ":I":
        lMode = _lIndicatif
    elif cMode == ":S":
        lMode = _lSubjonctif
    elif cMode.startswith((":I", ":S")):
        lMode = [cMode]
    else:
        return ""
    sWho = _dQuiEst.get(sSuj.lower(), None)
    if not sWho:
        if sSuj[0:1].islower(): # pas un pronom, ni un nom propre
            return ""
        sWho = ":3s"
    aSugg = set()
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        tTags = conj._getTags(sStem)
        if tTags:
            for sTense in lMode:
                if conj._hasConjWithTags(tTags, sTense, sWho):
                    aSugg.add(conj._getConjWithTags(sStem, tTags, sTense, sWho))
    if aSugg:
        return "|".join(aSugg)
    return ""


## Nouns and adjectives

def suggPlur (sFlex, sWordToAgree=None):
    "returns plural forms assuming sFlex is singular"
    if sWordToAgree:
        lMorph = _oSpellChecker.getMorph(sFlex)
        if sWordToAgree not in _dAnalyses and not _storeMorphFromFSA(sWordToAgree):
        if not lMorph:
            return ""
        sGender = cr.getGender(_dAnalyses.get(sWordToAgree, []))
        sGender = cr.getGender(lMorph)
        if sGender == ":m":
            return suggMasPlur(sFlex)
        elif sGender == ":f":
            return suggFemPlur(sFlex)
    aSugg = set()
    if "-" not in sFlex:
        if sFlex.endswith("l"):

    if aSugg:
        return "|".join(aSugg)
    return ""


def suggMasSing (sFlex, bSuggSimil=False):
    "returns masculine singular forms"
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    aSugg = set()
    for sMorph in _dAnalyses.get(sFlex, []):
    for sMorph in _oSpellChecker.getMorph(sFlex):
        if not ":V" in sMorph:
            # not a verb
            if ":m" in sMorph or ":e" in sMorph:
                aSugg.add(suggSing(sFlex))
            else:
                sStem = cr.getLemmaOfMorph(sMorph)
                if mfsp.isFemForm(sStem):

    if aSugg:
        return "|".join(aSugg)
    return ""


def suggMasPlur (sFlex, bSuggSimil=False):
    "returns masculine plural forms"
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    aSugg = set()
    for sMorph in _dAnalyses.get(sFlex, []):
    for sMorph in _oSpellChecker.getMorph(sFlex):
        if not ":V" in sMorph:
            # not a verb
            if ":m" in sMorph or ":e" in sMorph:
                aSugg.add(suggPlur(sFlex))
            else:
                sStem = cr.getLemmaOfMorph(sMorph)
                if mfsp.isFemForm(sStem):

    if aSugg:
        return "|".join(aSugg)
    return ""


def suggFemSing (sFlex, bSuggSimil=False):
    "returns feminine singular forms"
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    aSugg = set()
    for sMorph in _dAnalyses.get(sFlex, []):
    for sMorph in _oSpellChecker.getMorph(sFlex):
        if not ":V" in sMorph:
            # not a verb
            if ":f" in sMorph or ":e" in sMorph:
                aSugg.add(suggSing(sFlex))
            else:
                sStem = cr.getLemmaOfMorph(sMorph)
                if mfsp.isFemForm(sStem):

    if aSugg:
        return "|".join(aSugg)
    return ""


def suggFemPlur (sFlex, bSuggSimil=False):
    "returns feminine plural forms"
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    aSugg = set()
    for sMorph in _dAnalyses.get(sFlex, []):
    for sMorph in _oSpellChecker.getMorph(sFlex):
        if not ":V" in sMorph:
            # not a verb
            if ":f" in sMorph or ":e" in sMorph:
                aSugg.add(suggPlur(sFlex))
            else:
                sStem = cr.getLemmaOfMorph(sMorph)
                if mfsp.isFemForm(sStem):

            aSugg.add(e)
    if aSugg:
        return "|".join(aSugg)
    return ""


def hasFemForm (sFlex):
    "return True if there is a feminine form of <sFlex>"
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        if mfsp.isFemForm(sStem) or conj.hasConj(sStem, ":PQ", ":Q3"):
            return True
    if phonet.hasSimil(sFlex, ":f"):
        return True
    return False


def hasMasForm (sFlex):
    "return True if there is a masculine form of <sFlex>"
    for sStem in stem(sFlex):
    for sStem in _oSpellChecker.getLemma(sFlex):
        if mfsp.isFemForm(sStem) or conj.hasConj(sStem, ":PQ", ":Q1"):
            # what has a feminine form also has a masculine form
            return True
    if phonet.hasSimil(sFlex, ":m"):
        return True
    return False


def switchGender (sFlex, bPlur=None):
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    "return feminine or masculine form(s) of <sFlex>"
    aSugg = set()
    if bPlur == None:
        for sMorph in _dAnalyses.get(sFlex, []):
        for sMorph in _oSpellChecker.getMorph(sFlex):
            if ":f" in sMorph:
                if ":s" in sMorph:
                    aSugg.add(suggMasSing(sFlex))
                elif ":p" in sMorph:
                    aSugg.add(suggMasPlur(sFlex))
            elif ":m" in sMorph:
                if ":s" in sMorph:
                    aSugg.add(suggFemSing(sFlex))
                elif ":p" in sMorph:
                    aSugg.add(suggFemPlur(sFlex))
                else:
                    aSugg.add(suggFemSing(sFlex))
                    aSugg.add(suggFemPlur(sFlex))
    elif bPlur:
        for sMorph in _dAnalyses.get(sFlex, []):
        for sMorph in _oSpellChecker.getMorph(sFlex):
            if ":f" in sMorph:
                aSugg.add(suggMasPlur(sFlex))
            elif ":m" in sMorph:
                aSugg.add(suggFemPlur(sFlex))
    else:
        for sMorph in _dAnalyses.get(sFlex, []):
        for sMorph in _oSpellChecker.getMorph(sFlex):
            if ":f" in sMorph:
                aSugg.add(suggMasSing(sFlex))
            elif ":m" in sMorph:
                aSugg.add(suggFemSing(sFlex))
    if aSugg:
        return "|".join(aSugg)
    return ""


def switchPlural (sFlex):
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    "return plural or singular form(s) of <sFlex>"
    aSugg = set()
    for sMorph in _dAnalyses.get(sFlex, []):
    for sMorph in _oSpellChecker.getMorph(sFlex):
        if ":s" in sMorph:
            aSugg.add(suggPlur(sFlex))
        elif ":p" in sMorph:
            aSugg.add(suggSing(sFlex))
    if aSugg:
        return "|".join(aSugg)
    return ""


def hasSimil (sWord, sPattern=None):
    "return True if there is words phonetically similar to <sWord> (according to <sPattern> if required)"
    return phonet.hasSimil(sWord, sPattern)


def suggSimil (sWord, sPattern=None, bSubst=False):
    "return list of words phonetically similar to sWord and whom POS is matching sPattern"
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    aSugg = phonet.selectSimil(sWord, sPattern)
    for sMorph in _dAnalyses.get(sWord, []):
    for sMorph in _oSpellChecker.getMorph(sWord):
        aSugg.update(conj.getSimil(sWord, sMorph, bSubst))
        break
    if aSugg:
        return "|".join(aSugg)
    return ""


def suggCeOrCet (sWord):
    "suggest “ce” or “cet” or both according to the first letter of <sWord>"
    if re.match("(?i)[aeéèêiouyâîï]", sWord):
        return "cet"
    if sWord[0:1] == "h" or sWord[0:1] == "H":
        return "ce|cet"
    return "ce"


def suggLesLa (sWord):
    # we don’t check if word exists in _dAnalyses, for it is assumed it has been done before
    if any( ":p" in sMorph  for sMorph in _dAnalyses.get(sWord, []) ):
    "suggest “les” or “la” according to <sWord>"
    if any( ":p" in sMorph  for sMorph in _oSpellChecker.getMorph(sWord) ):
        return "les|la"
    return "la"


_zBinary = re.compile("^[01]+$")

def formatNumber (s):
    "add spaces or hyphens to big numbers"
    nLen = len(s)
    if nLen < 4:
        return s
    sRes = ""
    # nombre ordinaire
    nEnd = nLen
    while nEnd > 0:

    elif nLen == 9 and s.startswith("0"):
        sRes += "|" + s[0:3] + " " + s[3:5] + " " + s[5:7] + " " + s[7:9]                   # fixe belge 1
        sRes += "|" + s[0:2] + " " + s[2:5] + " " + s[5:7] + " " + s[7:9]                   # fixe belge 2
    return sRes


def formatNF (s):
    "typography: format NF reference (norme française)"
    try:
        m = re.match("NF[  -]?(C|E|P|Q|S|X|Z|EN(?:[  -]ISO|))[  -]?([0-9]+(?:[/‑-][0-9]+|))", s)
        if not m:
            return ""
        return "NF " + m.group(1).upper().replace(" ", " ").replace("-", " ") + " " + m.group(2).replace("/", "‑").replace("-", "‑")
    except:
        traceback.print_exc()
        return "# erreur #"


def undoLigature (c):
    "typography: split ligature character <c> in several chars"
    if c == "fi":
        return "fi"
    elif c == "fl":
        return "fl"
    elif c == "ff":
        return "ff"
    elif c == "ffi":

_xNormalizedCharsForInclusiveWriting = str.maketrans({
    '(': '_',  ')': '_',
    '.': '_',  '·': '_',
    '–': '_',  '—': '_',
    '/': '_'
 })
})


def normalizeInclusiveWriting (sToken):
    "typography: replace word separators used in inclusive writing by underscore (_)"
    return sToken.translate(_xNormalizedCharsForInclusiveWriting)

"""
# Grammalecte - Lexicographe
Grammalecte - Lexicographe
"""

# License: MPL 2


import re
import traceback


_dTAGS = {  
_dTAGS = {
    ':N': (" nom,", "Nom"),
    ':A': (" adjectif,", "Adjectif"),
    ':M1': (" prénom,", "Prénom"),
    ':M2': (" patronyme,", "Patronyme, matronyme, nom de famille…"),
    ':MP': (" nom propre,", "Nom propre"),
    ':W': (" adverbe,", "Adverbe"),
    ':J': (" interjection,", "Interjection"),

    ':O2': (" 2ᵉ pers.,", "Pronom : 2ᵉ personne"),
    ':O3': (" 3ᵉ pers.,", "Pronom : 3ᵉ personne"),
    ':C': (" conjonction,", "Conjonction"),
    ':Ĉ': (" conjonction (él.),", "Conjonction (élément)"),
    ':Cc': (" conjonction de coordination,", "Conjonction de coordination"),
    ':Cs': (" conjonction de subordination,", "Conjonction de subordination"),
    ':Ĉs': (" conjonction de subordination (él.),", "Conjonction de subordination (élément)"),
    

    ':Ñ': (" locution nominale (él.),", "Locution nominale (élément)"),
    ':Â': (" locution adjectivale (él.),", "Locution adjectivale (élément)"),
    ':Ṽ': (" locution verbale (él.),", "Locution verbale (élément)"),
    ':Ŵ': (" locution adverbiale (él.),", "Locution adverbiale (élément)"),
    ':Ŕ': (" locution prépositive (él.),", "Locution prépositive (élément)"),
    ':Ĵ': (" locution interjective (él.),", "Locution interjective (élément)"),

    'il': " pronom personnel sujet, 3ᵉ pers. masc. sing.",
    'on': " pronom personnel sujet, 3ᵉ pers. sing. ou plur.",
    'elle': " pronom personnel sujet, 3ᵉ pers. fém. sing.",
    'nous': " pronom personnel sujet/objet, 1ʳᵉ pers. plur.",
    'vous': " pronom personnel sujet/objet, 2ᵉ pers. plur.",
    'ils': " pronom personnel sujet, 3ᵉ pers. masc. plur.",
    'elles': " pronom personnel sujet, 3ᵉ pers. masc. plur.",
    

    "là": " particule démonstrative",
    "ci": " particule démonstrative",
    

    'le': " COD, masc. sing.",
    'la': " COD, fém. sing.",
    'les': " COD, plur.",
        

    'moi': " COI (à moi), sing.",
    'toi': " COI (à toi), sing.",
    'lui': " COI (à lui ou à elle), sing.",
    'nous2': " COI (à nous), plur.",
    'vous2': " COI (à vous), plur.",
    'leur': " COI (à eux ou à elles), plur.",

    'y': " pronom adverbial",
    "m'y": " (me) pronom personnel objet + (y) pronom adverbial",
    "t'y": " (te) pronom personnel objet + (y) pronom adverbial",
    "s'y": " (se) pronom personnel objet + (y) pronom adverbial",

    'en': " pronom adverbial",
    "m'en": " (me) pronom personnel objet + (en) pronom adverbial",
    "t'en": " (te) pronom personnel objet + (en) pronom adverbial",
    "s'en": " (se) pronom personnel objet + (en) pronom adverbial",
}


class Lexicographe:
    "Lexicographer - word analyzer"

    def __init__ (self, oSpellChecker):
        self.oSpellChecker = oSpellChecker
        self._zElidedPrefix = re.compile("(?i)^([dljmtsncç]|quoiqu|lorsqu|jusqu|puisqu|qu)['’](.+)")
        self._zCompoundWord = re.compile("(?i)(\\w+)-((?:les?|la)-(?:moi|toi|lui|[nv]ous|leur)|t-(?:il|elle|on)|y|en|[mts][’'](?:y|en)|les?|l[aà]|[mt]oi|leur|lui|je|tu|ils?|elles?|on|[nv]ous)$")
        self._zTag = re.compile("[:;/][\\w*][^:;/]*")

    def analyzeWord (self, sWord):
        "returns a tuple (a list of morphologies, a set of verb at infinitive form)"
        try:
            if not sWord:
                return (None, None)
            if sWord.count("-") > 4:
                return (["élément complexe indéterminé"], None)
            if sWord.isdigit():
                return (["nombre"], None)

                aMorph.append( "{} : {}".format(sWord, self.formatTags(lMorph[0])) )
            else:
                aMorph.append( "{} :  inconnu du dictionnaire".format(sWord) )
            # suffixe d’un mot composé
            if m2:
                aMorph.append( "-{} : {}".format(m2.group(2), self._formatSuffix(m2.group(2).lower())) )
            # Verbes
            aVerb = set([ s[1:s.find(" ")]  for s in lMorph  if ":V" in s ])
            aVerb = set([ s[1:s.find("/")]  for s in lMorph  if ":V" in s ])
            return (aMorph, aVerb)
        except:
            traceback.print_exc()
            return (["#erreur"], None)

    def formatTags (self, sTags):
        "returns string: readable tags"
        sRes = ""
        sTags = re.sub("(?<=V[1-3])[itpqnmr_eaxz]+", "", sTags)
        sTags = re.sub("(?<=V0[ea])[itpqnmr_eaxz]+", "", sTags)
        for m in self._zTag.finditer(sTags):
            sRes += _dTAGS.get(m.group(0), " [{}]".format(m.group(0)))[0]
        if sRes.startswith(" verbe") and not sRes.endswith("infinitif"):
            sRes += " [{}]".format(sTags[1:sTags.find(" ")])

"""
# Masculins, féminins, singuliers et pluriels
Masculins, féminins, singuliers et pluriels
"""

from .mfsp_data import lTagMiscPlur as _lTagMiscPlur
from .mfsp_data import lTagMasForm as _lTagMasForm
from .mfsp_data import dMiscPlur as _dMiscPlur
from .mfsp_data import dMasForm as _dMasForm

"""
# Grammalecte - Suggestion phonétique
Grammalecte - Suggestion phonétique
"""

# License: GPL 3

import re

from .phonet_data import dWord as _dWord
from .phonet_data import lSet as _lSet
from .phonet_data import dMorph as _dMorph

#! python3
# coding: UTF-8

"""
Grammar checker tests for French language
"""

import unittest
import os
import re
import time

#!python3

"""
Text formatter
"""

import re


dReplTable = {
    # surnumerary_spaces
    "start_of_paragraph":          [("^[  ]+", "")],

    "erase_non_breaking_hyphens":  [("­", "")],
    ## typographic signs
    "ts_apostrophe":          [ ("(?i)\\b([ldnjmtscç])['´‘′`](?=\\w)", "\\1’"),
                                ("(?i)(qu|jusqu|lorsqu|puisqu|quoiqu|quelqu|presqu|entr|aujourd|prud)['´‘′`]", "\\1’") ],
    "ts_ellipsis":            [ ("\\.\\.\\.", "…"),
                                ("(?<=…)[.][.]", "…"),
                                ("…[.](?![.])", "…") ],
    "ts_n_dash_middle":       [ (" [-—] ", " – "), 
    "ts_n_dash_middle":       [ (" [-—] ", " – "),
                                (" [-—],", " –,") ],
    "ts_m_dash_middle":       [ (" [-–] ", " — "),
                                (" [-–],", " —,") ],
    "ts_n_dash_start":        [ ("^[-—][  ]", "– "),
                                ("^– ", "– "),
                                ("^[-–—](?=[\\w.…])", "– ") ],
    "ts_m_dash_start":        [ ("^[-–][  ]", "— "),

# ERREURS COURANTES
# http://fr.wikipedia.org/wiki/Wikip%C3%A9dia:Fautes_d%27orthographe/Courantes


!!
!!
!! Options                                                                                          
!! Options                                                                                        !!
!!
!!

OPTGROUP/basic: typo apos, esp tab, nbsp unit, tu maj, num virg, nf chim, ocr mapos, liga
OPTGROUP/gramm: conf sgpl gn
OPTGROUP/verbs: infi conj ppas, imp inte vmode
OPTGROUP/style: bs pleo, redon1 redon2, neg

OPTLABEL/debug:     Debug
OPTLABEL/idrule:    Display control rule identifier [!]|Display control rule identifier in the context menu message.



!!
!!
!! Définitions pour les regex                                                                       
!! Définitions pour les regex                                                                     !!
!!
!!

# REGEX
DEF: avoir          [aeo]\w*
DEF: etre           [êeésf]\w+
DEF: avoir_etre     [aeêésfo]\w*
DEF: aller          (?:ai?ll|v[ao]|ir[aio])\w*
DEF: ppas           \w[\w-]+[éiust]e?s?
DEF: infi           \w[\w-]+(?:er|ir|re)
DEF: w_1            \w[\w-]*
DEF: w_2            \w[\w-]+
DEF: w_3            \w[\w-]+\w
DEF: w_4            \w[\w-][\w-]+\w
DEF: w1             \w+
DEF: w2             \w\w+
DEF: w3             \w\w\w+
DEF: w4             \w\w\w\w+


# GRAPH
DEF: mois           [>janvier|>février|>mars|>avril|>mai|>juin|>juillet|>août|>aout|>septembre|>octobre|>novembre|>décembre|>vendémiaire|>brumaire|>frimaire|>nivôse|>pluviôse|>ventôse|>germinal|>floréal|>prairial|>messidor|>thermidor|>fructidor]
DEF: mi_mois        [>mi-janvier|>mi-février|>mi-mars|>mi-avril|>mi-mai|>mi-juin|>mi-juillet|>mi-août|>mi-aout|>mi-septembre|>mi-octobre|>mi-novembre|>mi-décembre|mi-vendémiaire|mi-brumaire|mi-frimaire|mi-nivôse|mi-pluviôse|mi-ventôse|mi-germinal|mi-floréal|mi-prairial|mi-messidor|mi-thermidor|mi-fructidor]
DEF: pronom_obj     [moi|toi|soi|lui|elle|nous|vous|eux|elles|moi-même|toi-même|soi-même|lui-même|elle-même|nous-mêmes|vous-même|vous-mêmes|eux-mêmes|elles-mêmes]


!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!! PASSE 0: PARAGRAPHE PAR PARAGRAPHE                                                               
!! PASSE 0: PARAGRAPHE PAR PARAGRAPHE                                                             !!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!
!!



!!
!!
!!! Espaces & tabulations                                                                           
!!! Espaces & tabulations                                                                         !!
!!
!!

# Espaces surnuméraires
# Note : les tabulations ne sont pas soulignées dans LibreOffice. Mais l’erreur est bien présente.
__<s>/tab(tab_début_ligne)__           ^[      ]+ <<- ->> ""                   # Espace(s) en début de ligne à supprimer : utilisez les retraits de paragraphe.
__<s>/tab(tab_fin_ligne)__             [       ]+$ <<- ->> ""                  # Espace(s) en fin de ligne à supprimer.

# À partir d’ici, toute règle est susceptible de supprimer des caractères et les remplacer par des espaces ou des chaînes de longueur égale.
#
#


!!!
!!!
!!! Processeur: efface les ponctuations gênantes (URL, sigles, abréviations, IP, heures, etc.)      
!!! Processeur: efface les ponctuations gênantes (URL, sigles, abréviations, IP, heures, etc.)    !!
!!!
!!!

# e-mail
__<i>(p_email)__
    \w[\w.-]*@\w[\w.-]*\w[.]\w+ <<- ~>> *

# URL
__<i>(p_URL)__
    https?://[\w./?&!%=+*"'@$#-]+ <<- ~>> *
__<i](p_URL2)__
    ((?:{w_1}[.])*)({w_2})([.](?:com|net|org|info|fr|ca|be|ch|i[ot]|co[.]uk|tk|es|jp|zh|ru|us|nl|xyz)) @@0,**,$
    <<- ~1>> *
    <<- ~2>> =\2.capitalize()
    <<- =>> define(\2, [":MP:e:i"])
    <<- ~3>> *

# Numéro de chapitre
__<i>(p_chapitre)__
    ^\d+[.][\d.-]* <<- ~>> *

# Numéro suivi de plusieurs espaces, considéré comme une numérotation de chapitre

TEST: Dossier C:\Program Files (x86)\LibreOffice
TEST: [1] Dossier à revoir.



!!!
!!!
!!! Processeur: balises HTML et LaTeX                                                               
!!! Processeur: balises HTML et LaTeX                                                             !!
!!!
!!!

# HTML
__<i>/html(p_html_amp_xxx)__            &amp;[a-zA-Z]+; <<- ~>> _
__<i>/html(p_html_lt)__                 &lt; <<- ~>> "   <"
__<i>/html(p_html_gt)__                 &gt; <<- ~>> >

TEST: __html__  des <b>{{homme}}</b>
TEST: __html__  des [b]{{femme}}[/b]



!!
!!
!!!! Écritures épicènes dystypographiques                                                           
!!!! Écritures épicènes dystypographiques                                                         !!
!!
!!

# (attention aux modifs: brainfuck d’intensité non négligeable)
# La désambiguïsation est faite lors de la deuxième passe

__[i](typo_écriture_épicène_tous_toutes)__

TEST: je suis {{déconsidéré.e}} par ma hiérarchie.
TEST: il faut en parler à l’{{auteur(e)}} et à son agent.


!!
!!
!!!! Majuscules manquantes                                                                          
!!!! Majuscules manquantes                                                                        !!
!!
!!

# Majuscules après un point
__[s]/maj(majuscule_après_point)__
    ({w_2})[.] ([a-zéàèîçô]\w*)  @@0,$
    <<- not re.search("(?i)^(?:etc|[A-Z]|chap|cf|fig|hab|litt|circ|coll|r[eé]f|étym|suppl|bibl|bibliogr|cit|op|vol|déc|nov|oct|janv|juil|avr|sept)$", \1)

    <<- after("\w\w[.] +\w+") -1>> =\1.capitalize()                                                 # Majuscule en début de phrase, sauf éventuellement lors d’une énumération.

TEST: {{je}} suis disponible quand tu veux. Mais pas aujourd’hui.


!!
!!
!!!! Virgules                                                                                       
!!!! Virgules                                                                                     !!
!!
!!

# virgules manquantes
__[i>/virg(virgule_manquante_avant_etc)__   {w_1}( etc[.])  @@$ <<- -1>> , etc.                     # Avant « etc. », il faut mettre une virgule.
__[i>/virg(virgule_manquante_avant_car)__
    ({w_1})( car)(?= (?:j[e’]|tu|ils?|nous|vous|elles?|on|les?|l[a’]|ces?|des?|cette|[mts](?:on|a|es))\b)  @@0,$
    <<- not morph(\1, ":[DR]", False) -2>> , car
    # Si « car » est la conjonction de coordination, une virgule est peut-être souhaitable.|http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?id=3447
__[i>/virg(virgule_manquante_avant_mais)__
    ({w_1})( mais)(?= (?:j[e’]|tu|ils?|nous|vous|elles?|on)\b)  @@0,$
    <<- not morph(\1, ">(?:[mtscl]es|[nv]os|quels) ", False) -2>> , mais
    <<- not morph(\1, ">(?:[mtscl]es|[nv]os|quels)/", False) -2>> , mais
    # Si « mais » est la conjonction de coordination, une virgule est souhaitable si elle introduit une nouvelle proposition.|http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?id=3445
__[i>/virg(virgule_manquante_avant_donc)__
    ({w_1})( donc)(?= (?:j[e’]|tu|ils?|elles?|on)\b)  @@0,$
    <<- not morph(\1, ":V", False) -2>> , donc
    # Si « mais » est la conjonction de coordination, une virgule est souhaitable si elle introduit une nouvelle proposition.|http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?id=3448

TEST: Un chien, un chat{{ etc.}}                                            ->> , etc.

#__<i>/typo__  ({w_2}) *$   @@0
#    <<- before("[a-zéàùèê][.] ") and not before(r"^(?:\d|\w[.])")
#    -1>> \1.|\1…|\1 !|\1 ?|\1 :|\1 ;|\1,                                               # Il manque une ponctuation finale.


!!
!!
!!!! Espaces manquants                                                                              
!!!! Espaces manquants                                                                            !!
!!
!!

__<s>/typo(typo_espace_manquant_après1)__
    ({w1})[,:]({w_1})  @@0,$  <<- not \1.isdigit() -2>> " \2"                                       # Il manque un espace.
__<s>/typo(typo_espace_manquant_après2)__
    [?!;%‰‱˚»}]({w_1})  @@$  <<- -1>> " \1"                                                       # Il manque un espace.

TEST: espace.{{Manquant}}                                                   ->> " Manquant"
TEST: qui ont refuse.{{de}} se sauver                                       ->> " de"
TEST: pic.twitter.com/PICNAME


!!
!!
!!!! Points                                                                                         
!!!! Points                                                                                       !!
!!
!!

# Points superflus
__[i>/typo(typo_et_cetera)__             etc(?:[.]{3,5}|…) <<- ->> etc.                             # Un seul point après « etc. »

TEST: {{etc…}}

TEST: Ligne de points{{...........................}}
TEST: Aujourd’hui, M. Camus va nous lire son texte.



!!
!!
!!!! Tirets                                                                                         
!!!! Tirets                                                                                       !!
!!
!!

# Tirets, énumérations et dialogues
__<s>/typo(typo_tiret_début_ligne)__    ^[-_][  ] <<- ->> "— |– "           # Dialogues et énumérations : un tiret cadratin ou demi-cadratin, suivi d’un espace insécable, est requis.
__<s>/typo(typo_tiret_dans_dialogue)__  ^ *«[  ](-[  ]) @@$ <<- -1>> "— |– "
    # Dialogues : un tiret cadratin ou demi-cadratin, suivi d’un espace insécable, est requis. (Attention : à moins qu’il s’agisse d’un dialogue inclus dans un autre dialogue, cette manière d’écrire est erronée.)

TEST: {{—}}Viens là.                                                        ->> "— |– "
TEST: Cette phrase contient{{ - }}c’est certain – des tirets d’incise.      ->> " – | — "
TEST: « {{- }}Viens ! On va en finir avec cette affaire, sale fils de pute. ->> "— |– "


!!
!!
!!!! Ponctuations redondantes                                                                       
!!!! Ponctuations redondantes                                                                     !!
!!
!!

# virgules et points
__<s>/typo(typo_virgules_points)__ 
__<s>/typo(typo_virgules_points)__
    ,[.,]{2,}
    <<- ->> =\0.replace(",", ".").replace("...", "…")                                               # Erreur de numérisation ? Virgules au lieu de points ?

TEST: Ah !{{,.,}} et en quoi consistait le festin ?


__<s>/typo(typo_ponctuation_superflue1)__    , ?([.,;:!?]) @@$ <<- ->> =",|" + \1       # Une de ces ponctuations est superflue.

TEST: Succession : a{{;}} b.                                                ->> " ;"
TEST: C’est assez {{: }}parlez !                                            ->> ": "
TEST: {{» }}C’est ce qu’on croit savoir.                                    ->> "» |« "


!!
!!
!!!! Signes typographiques                                                                          
!!!! Signes typographiques                                                                        !!
!!
!!

__<s>/typo(typo_signe_multiplication)__
    (\d+) ?[x*] ?(\d+)  @@0,$ <<- not \0.startswith("0x") ->> \1 × \2                               # Signe de multiplication typographique.

TEST: {{800 x 600}}                                                         ->> 800 × 600

TEST: __liga__ un {{fl}}ottement                                           ->> fl
TEST: __liga__ une belle {{fi}}gure                                        ->> fi


!!
!!
!!!! Apostrophes                                                                                    
!!!! Apostrophes                                                                                  !!
!!
!!
## Apostrophe typographique
__[i>/apos(apostrophe_typographique)__
    ([ldsncjmç]|jusqu|lorsqu|aujourd|presqu|quelqu|puisqu|qu|prud|entr)['´‘′`](?=[\w"«]) @@0 <<- ->> \1’    # Apostrophe typographique.
__[i>/apos(apostrophe_typographique_après_t)__
    t['´‘′`](?!ils?|elles?|on)(?=[\w"«]) <<- ->> t’                                                         # Apostrophe typographique.

TEST: Il {{s }}était dit qu’il ne le ferait pas.                                ->> s’
TEST: Je {{m }}étais imaginé que je pourrais pas retrouver cet objet.           ->> m’
TEST: Je {{n }}ai pas retrouvé l’ambiance de mes années de lycée.               ->> n’


!!
!!
!!!! Guillemets typographiques                                                                      
!!!! Guillemets typographiques                                                                    !!
!!
!!

__<s>/typo(typo_guillemets_typographiques_doubles_ouvrants)__
    "(?=\w)
    <<- not before(r"[a-zA-Zéïîùàâäôö]$") ->> « |“                                                  # Guillemets typographiques ouvrants.
__<s>/typo(typo_guillemets_typographiques_doubles_fermants)__

TEST: Le {{'}}grand’ homme.                                                 ->> “|‘|‹
TEST: Le ‘grand{{'}} homme.                                                 ->> ”|’|›
TEST: J’en ai '''marre''' (syntaxe wiki).


!!
!!
!!!! Élisions                                                                                       
!!!! Élisions                                                                                     !!
!!
!!

__[i]/typo(typo_élision_déterminants)__
    (l[ea] |de )([aâeéêiîoôu]\w+)  @@0,$
    <<- not re.search("(?i)^(?:onz[ei]|énième|iourte|ouistiti|ouate|one-?step|ouf|Ouagadougou|I(?:I|V|X|er|ᵉʳ|ʳᵉ|è?re))", \2) and not \2.isupper() and not morph(\2, ":G", False)
    -1>> =\1[0]+"’"                                         # Élision de l’article devant un mot commençant par une voyelle.|http://fr.wikipedia.org/wiki/Élision

    -1>> cet                                                # Euphonie : « ce » devient « cet » devant un mot commençant par une voyelle.

TEST: {{ce}} animal est dangereux


!!
!!
!!!! Divers                                                                                         
!!!! Divers                                                                                       !!
!!
!!

## NF (memo: don’t use flag i)
__[s]/nf(nf_norme_française)__
    [nN][fF][  -]?(?:c|C|e|E|p|P|q|Q|s|S|x|X|z|Z|[eE][nN](?:[  -][iI][sS][oO]|))[  -]?[0-9]+(?:[/ ‑-][0-9]+|)
    <<- not re.search("^NF (?:C|E|P|Q|S|X|Z|EN(?: ISO|)) [0-9]+(?:‑[0-9]+|)", \0)

    <<- ->> =\0.replace("2", "₂").replace("3", "₃").replace("4", "₄")                               # Typographie des composés chimiques. [!]

TEST: __chim__ les molécules {{CaCO3}} et {{H2O}}…


!!
!!
!!!! Cohérence des guillemets                                                                       
!!!! Cohérence des guillemets                                                                     !!
!!
!!

__<s>/typo(typo_cohérence_guillemets_chevrons_ouvrants)__
    («)[^»“]+?(”)  @@0,$
    <<- not before(r"\w$") -1>> “                                                                   # Guillemet ouvrant différent du guillemet fermant (”).
    <<- not after(r"^\w") -2>> " »"                                                                 # Guillemet fermant différent du guillemet ouvrant («).

TEST: « Je suis là.
TEST: — Oui, nous t’avons vu arriver. »
TEST: « J’en suis “malade”. »


!!
!!
!!!! Espaces insécables avant unités de mesure                                                      
!!!! Espaces insécables avant unités de mesure                                                    !!
!!
!!

__[s]/unit(unit_nbsp_avant_unités1)__
    ((\d+(?:,\d+[⁰¹²³⁴⁵⁶⁷⁸⁹]?|[⁰¹²³⁴⁵⁶⁷⁸⁹]|)) ?)(?:[kcmµn]?(?:[slgJKΩ]|m[²³]?|Wh?|Hz|dB)|[%‰€$£¥Åℓhj]|min|°C|℃)(?![’'])  @@0,0
    <<- -1>> "\2 "
    # Avec une unité de mesure, mettez un espace insécable.

TEST: 1998 s’annonce grandiose.
TEST: il y en a 3,43 milliards.
TEST: je veux 200 euros.


!!
!!
!!!! Grands nombres                                                                                 
!!!! Grands nombres                                                                               !!
!!
!!

__[s]/num(num_grand_nombre_soudé)__
    \d\d\d\d+
    <<- not before("NF[  -]?(C|E|P|Q|X|Z|EN(?:[  -]ISO|)) *$") >>>
    <<- \0.__len__() > 4 ->> =formatNumber(\0)                                                      # Formatage des grands nombres.

    <<- ~>> =\0.replace(" ", "_")

TEST: Il a perdu {{20 000}} euros à la Bourse en un seul mois.


!!
!!
!!!! Dates                                                                                          
!!!! Dates                                                                                        !!
!!
!!
__[i]/date(date_nombres)__
    (?<!\d[ /.-])(\d\d?)[ /.-](\d\d?)[ /.-](\d\d\d+)(?![ /.-]\d)  @@0,w,$
    <<- not checkDate(\1, \2, \3) and not before(r"(?i)\bversions? +$") ->> _                       # Cette date est invalide.
    <<- ~>> =\0.replace(".", "-").replace(" ", "-").replace("\/", "-")

TEST: le {{29 02 2011}}
TEST: le {{40-02-2011}}
TEST: le {{32.03.2018}}
TEST: le {{81/01/2012}}
TEST: 12-12-2012


!!
!!
!!!! Redondances                                                                                    
!!!! Redondances                                                                                  !!
!!
!!
__[i]/redon1(redondances_paragraphe)__
    ({w_4})[  ,.;!?:].*[  ](\1)  @@0,$
    <<- not morph(\1, ":(?:G|V0)|>(?:t(?:antôt|emps|rès)|loin|souvent|parfois|quelquefois|côte|petit|même) ", False) and not \1[0].isupper()
    <<- not morph(\1, ":(?:G|V0)|>(?:t(?:antôt|emps|rès)|loin|souvent|parfois|quelquefois|côte|petit|même)/", False) and not \1[0].isupper()
    -2>> _                                                      # Dans ce paragraphe, répétition de « \1 » (à gauche).
    <<- __also__ -1>> _                                         # Dans ce paragraphe, répétition de « \1 » (à droite).

TEST: __redon1__ Tu es son {{avenir}}. Et lui aussi est ton {{avenir}}.
TEST: __redon1__ Car parfois il y en a. Mais parfois il n’y en a pas.




!!!
!!!
!!! Processeur: Dernier nettoyage avant coupure du paragraphe en phrases                            
!!! Processeur: Dernier nettoyage avant coupure du paragraphe en phrases                          !!
!!!
!!!

# Trait d’union conditionnel (u00AD)
__<i>(p_trait_union_conditionnel1)__    \w+‑\w+‑\w+ <<- ~>> =\0.replace("‑", "")
__<i>(p_trait_union_conditionnel2)__    \w+‑\w+ <<- ~>> =\0.replace("‑", "")

# empêcher la scission en fin de dialogue
__<s>(p_fin_dialogue1)__    ([?!…][?!…  ]*)[ "'”» ]*,  @@0 <<- ~1>> *
__<s>(p_fin_dialogue2)__    ([?!…][?!…  ]*)[ "'”» ]*[a-zéèêîô]  @@0 <<- ~1>> ,

TEST: « Je suis donc perdu ? », dit Paul.
TEST: “C’est bon !”, croit savoir Marie.
TEST: “Parce que… ?” finit par demander Paul.
TEST: « Dans quel pays sommes-nous ? » demanda un manifestant. 
TEST: « Dans quel pays sommes-nous ? » demanda un manifestant.


!!
!!
!!
!!
!!

!!
!!
!!
!!
!!
!!
!!
!! PASSE 1: PHRASE PAR PHRASE                                                                       
!! PASSE 1: PHRASE PAR PHRASE                                                                     !!
!!
!!
!!
!!
!!
!!
!!

!!
!!

[++]



!!!! Doublons (casse identique)                                                                     
!!!! Doublons (casse identique)                                                                   !!

__[s](doublon)__
    ({w1}) {1,3}\1  @@0
    <<- not re.search("(?i)^([nv]ous|faire|en|la|lui|donnant|œuvre|h[éoa]|hou|olé|joli|Bora|couvent|dément|sapiens|très|vroum|[0-9]+)$", \1)
        and not (re.search("^(?:est|une?)$", \1) and before("[’']$"))
        and not (\1 == "mieux" and before("(?i)qui +$"))
    ->> \1   # Doublon.

TEST: Il y a un {{doublon doublon}}.


!!!! Nombres: typographie                                                                           
!!!! Nombres: typographie                                                                         !!

#(\d\d\d\d)-(\d\d\d\d)   <<- ->> \1–\2                              # Ne pas séparer deux dates par un trait d’union, mais par un tiret demi-cadratin.

__[s]/num(num_lettre_O_zéro1)__  [\dO]+[O][\dO]+ <<- not option("ocr") ->> =\0.replace("O", "0")    # S’il s’agit d’un nombre, utilisez le chiffre « 0 » plutôt que la lettre « O ».
__[s]/num(num_lettre_O_zéro2)__  [1-9]O <<- not option("ocr") ->> =\0.replace("O", "0")             # S’il s’agit d’un nombre, utilisez le chiffre « 0 » plutôt que la lettre « O ».

TEST: année {{2O11}}                                                      ->> 2011

TEST: le {{IInd}} siècle                        ->> IIᵈ
TEST: le {{VIième}} siècle                      ->> VIᵉ
TEST: Le {{XXIème}} siècle.                     ->> XXIᵉ
TEST: le {{XXè}} siècle.                        ->> XXᵉ



!!!! Écritures épicènes invariables                                                                 
!!!! Écritures épicènes invariables                                                               !!

__[i](d_typo_écriture_épicène_pluriel)__
    ({w_1}[éuitsrn])_(?:[nt]|)e_s  @@0
    <<- morphex(\1, ":[NAQ]", ":G") =>> define(\1, [":N:A:Q:e:p"])

__[i](d_typo_écriture_épicène_singulier)__
    ({w_2}[éuitsrn])_e  @@0
    <<- morph(\1, ":[NAQ]", False) =>> define(\1, [":N:A:Q:e:s"])


!!!! Dates                                                                                          
!!!! Dates                                                                                        !!

__[i]/date(date_jour_mois_année)__
    (\d\d?) (janvier|février|ma(?:rs|i)|a(?:vril|o[ûu]t)|jui(?:n|llet)|septembre|octobre|novembre|décembre) (\d\d\d+)  @@0,w,$ 
    (\d\d?) (janvier|février|ma(?:rs|i)|a(?:vril|o[ûu]t)|jui(?:n|llet)|septembre|octobre|novembre|décembre) (\d\d\d+)  @@0,w,$
    <<- not checkDateWithString(\1, \2, \3) ->> _                                                   # Cette date est invalide.

TEST: {{29 février 2011}}


__[i]/date(date_journée_jour_mois_année1)__
    (lundi|m(?:ardi|ercredi)|jeudi|vendredi|samedi|dimanche),? (?:le |)(\d\d?)-(\d\d?)-(\d\d+)  @@0,w,w,$ 
    (lundi|m(?:ardi|ercredi)|jeudi|vendredi|samedi|dimanche),? (?:le |)(\d\d?)-(\d\d?)-(\d\d+)  @@0,w,w,$
    <<- not after(r"^ +av(?:ant|) +J(?:C|ésus-Christ)") and not checkDay(\1, \2, \3, \4)
    -1>> =getDay(\2, \3, \4)                                                                        # Le jour de la date suivante est incorrect.

TEST: {{mercredi}}, le 10-06-2014                                         ->> mardi


__[i]/date(date_journée_jour_mois_année2)__
    (lundi|m(?:ardi|ercredi)|jeudi|vendredi|samedi|dimanche),? (?:le |)(\d\d?) (janvier|février|ma(?:rs|i)|a(?:vril|o[ûu]t)|jui(?:n|llet)|septembre|octobre|novembre|décembre) (\d\d+)  @@0,w,w,$ 
    (lundi|m(?:ardi|ercredi)|jeudi|vendredi|samedi|dimanche),? (?:le |)(\d\d?) (janvier|février|ma(?:rs|i)|a(?:vril|o[ûu]t)|jui(?:n|llet)|septembre|octobre|novembre|décembre) (\d\d+)  @@0,w,w,$
    <<- not after(r"^ +av(?:ant|) +J(?:C|ésus-Christ)") and not checkDayWithString(\1, \2, \3, \4)
    -1>> =getDayWithString(\2, \3, \4)                                                              # Le jour de la date suivante est incorrect.

TEST: {{mercredi}}, le 10 juin 2014                                       ->> mardi
TEST: {{lundi}}, 18 août 1912                                             ->> dimanche
TEST: lundi, 18 août 1912 avant Jésus-Christ (date imaginaire)


__[i]/date(date_mois_31)__ 31 (avril|juin|septembre|novembre)  @@3
    <<- ->> 30 \1                                                                                   # Cette date est invalide. Il n’y a que 30 jours en \1. 
    <<- ->> 30 \1                                                                                   # Cette date est invalide. Il n’y a que 30 jours en \1.

TEST: le {{31 avril}}


__[i]/date(date_février)__ 3[01] février
    <<- ->> 28 février|29 février                                                                   # Cette date est invalide. Il n’y a que 28 ou 29 jours en février. 
    <<- ->> 28 février|29 février                                                                   # Cette date est invalide. Il n’y a que 28 ou 29 jours en février.

TEST: le {{30 février}}



!!!
!!!
!!! Processeur: épuration des signes inutiles et quelques simplifications                           
!!! Processeur: épuration des signes inutiles et quelques simplifications                         !!
!!!
!!!

# fin de phrase
__<s>(p_fin_de_phrase)__        [.?!:;…][ .?!… »”")]*$  <<- ~>> *

# début de phrase

__[i](p_références_aux_notes)__ [a-zéèâàôîù][a-zéèâàôîù-]+(\d+) @@$ <<- not morph(\0, ":", False) ~1>> *

# faux positifs avec adverbes de négation
__[i](p_pas_mal)__              pas mal <<- not morph(word(-1), ":D", False) ~>> *
__[i](p_pas_assez)__            pas assez ({w_2}) @@$ <<- morph(\1, ":A", False) and not morph(word(-1), ":D", False) ~>> *

# faux positifs avec «à chez»
__[i](p_de_chez_à_chez_pronom)__    de chez \w+ (?:à|jusqu à) chez (?:moi|toi|lui|elles?|eux|nous|vous) <<- ~>> * 
__[i](p_de_chez_à_chez_pronom)__    de chez \w+ (?:à|jusqu à) chez (?:moi|toi|lui|elles?|eux|nous|vous) <<- ~>> *
__[i](p_de_chez)__                  (jusqu à|de) chez @@0 <<- ~1>> *

# faux positifs
__[i](p_en_tout_et_pour_tout)__ en tout et pour tout <<- ~>> *
__[i](p_au_sortir_de)__         au (sortir) de?s?  @@3 <<- ~1>> *
__[i](p_au_revoir)__            au revoir <<- ~>> au_revoir

TEST: New York {{étaient}} {{devenue}} la plaque tournante de tous les trafics.


!!
!!
!!!! Traits d’union                                                                                 
!!!! Traits d’union                                                                               !!
!!
!!

__<i]/tu(tu_t_euphonique_incorrect)__
    ([-–—− ]t(?:[’' ][-–—−]?|[-–—−][’' ]?))(ils?|elles?|on|tu)  @@0,$
    <<- re.search("(?i)^(?:ils|elles|tu)$", \2) -1>> -      # Le “t” euphonique n’est pas nécessaire avec “\2”.|http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?T1=t+euphonique&id=2513
    <<- __else__ and \1 != "-t-" and \1 != "-T-" -1>> -t-   # Pour le “t” euphonique, il faut deux traits d’union. Pas d’apostrophe. Pas d’espace.
    <<- ~1>> -t-
    <<- \1 != "-t-" ~1>> -t-
__<i]/tu(tu_t_euphonique_superflu)__
    [td]([- ]t[-’' ])(?:il|elle|on)  @@1
    <<- -1>> -                                              # Le “t” euphonique est superflu quand le verbe se termine par “t” ou “d”.|http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?T1=t+euphonique&id=2513
    <<- ~1>> -t-
    <<- \1 != "-t-" ~1>> -t-
__<i]/tu(tu_t_euphonique_manquant)__
    [aec](-(il|elle|on))  @@1,2  <<- -1>> -t-\2             # Il faut un “t” euphonique.|http://bdl.oqlf.gouv.qc.ca/bdl/gabarit_bdl.asp?T1=t+euphonique&id=2513

TEST: va{{ t’}}il y parvenir ?                          ->> -t-
TEST: A{{ t’}}elle soif ?                               ->> -t-
TEST: A{{ t-}}elle faim ?                               ->> -t-
TEST: a{{ t'}}elle                                      ->> -t-

TEST: Cet {{in folio}} est incroyable.
TEST: c’est {{in}} stage de réinsertion.
TEST: Dans ce drive in douze hommes mangent.


__[i]/tu(tu_préfixe_mi)__
    mi ({w2}) @@$  <<- morph(\1, ":[NAQ]", False) ->> mi-\1                                         # S’il s’agit d’un seul mot, il manque un trait d’union.    
    mi ({w2}) @@$  <<- morph(\1, ":[NAQ]", False) ->> mi-\1                                         # S’il s’agit d’un seul mot, il manque un trait d’union.

TEST: J’ai été engagé pour un {{mi temps}}.


__[i]/tu(tu_préfixe_quasi)__
    (?:l(?:es?|a|eurs?)|ce(?:tte|t|s|)|des?|m(?:a|on|es)|[ts](?:es|a)) (quasi ({w1}))  @@$,$ 
    (?:l(?:es?|a|eurs?)|ce(?:tte|t|s|)|des?|m(?:a|on|es)|[ts](?:es|a)) (quasi ({w1}))  @@$,$
    <<- morphex(\2, ":N", ":[AGW]") -1>> quasi-\2                                                   # Il manque un trait d’union : « quasi » s’accroche au mot qui le suit s’il s’agit d’un nom.
    <<- ~1>> \2

TEST: leurs {{quasi indifférences}}


__[i]/tu(tu_préfixe_semi)__

__[i]/tu(tu_là_adverbe)__       là (bas|haut|desso?us|dedans|devant|derrière) @@3 <<- ->> là-\1     # Il manque un trait d’union.
__[i]/tu(tu_n_est_ce_pas)__             n’est(?: ce[ -]pas|-ce-pas) <<- ->> n’est-ce pas            # Un trait d’union.
__[i]/tu(tu_nord_sud_ouest)__           (nord|sud) ouest @@0 <<- ->> \1-ouest                       # Il manque un trait d’union.
__[i]/tu(tu_nord_sud_est)__             (nord|sud) est @@0 <<- isEnd() ->> \1-est                   # Il manque un trait d’union.
__[i]/tu(tu_outre_mer)__                outre mer <<- ->> outre-mer                                 # Il manque un trait d’union.
__[i]/tu(tu_ouï_dire)__                 ou[iï] dire <<- morph(word(-1), ":G") ->> ouï-dire          # Il manque un trait d’union.
__[i]/tu(tu_par_préposition)__
    par (desso?us|devant|delà|derrière|dehors|dedans|devers)  @@$ 
    par (desso?us|devant|delà|derrière|dehors|dedans|devers)  @@$
    <<- ->> par-\1                                                                                  # Il manque un trait d’union.
__[i]/tu(tu_par_ci_par_là)__            par ci,? par là <<- ->> par-ci par-là|par-ci, par-là        # Trait(s) d’union manquant(s).
__[i]/tu(tu_prêt_à_porter)__
    (prêts?) à porter  @@0 
    (prêts?) à porter  @@0
    <<- before(r"(?i)\b(?:les?|du|des|un|ces?|[mts]on) +") ->> \1-à-porter                          # Il manque les traits d’union.
__[i]/tu(tu_plate_forme)__  plates? formes? <<- ->> plate-forme|plates-formes|plateforme|plateformes    # Il manque un trait d’union. Vous pouvez aussi souder les deux mots.
__[i]/tu(tu_quelques_uns_unes)__       quelques (une?s) @@$ <<- ->> quelques-\1                     # Il manque un trait d’union.
__[i]/tu(tu_plus_moins_values)__       (plus|moins) (values?)  @@0,$  <<- ->> \1-\2                 # Il manque un trait d’union.
__[i]/tu(tu_rez_de_chaussée)__         rez de chaussées? <<- ->> rez-de-chaussée                    # Il manque un trait d’union.
__[i]/tu(tu_science_fiction)__         science fiction <<- ->> science-fiction                      # Il manque un trait d’union.
__[i]/tu(tu_stock_option)__            stock options? <<- ->> =\0.replace(" ", "-")                 # Il manque un trait d’union.

TEST: on a fait {{à la va vite}}…
TEST: marre de faire des {{allers retours}}
TEST: {{le}} {{bouche à oreille}}
TEST: à ce {{moment là}}                                                        ->> moment-là
TEST: une {{plus value}}                                                        ->> plus-value
TEST: Il est {{en-dessous}} de tout.                                            ->> en dessous
TEST: Ils sont {{en-deçà}} de tout                                              ->> en deçà
TEST: {{Là bas}}.                                                               ->> Là-bas 
TEST: {{Là bas}}.                                                               ->> Là-bas
TEST: {{Au dessus}}                                                             ->> Au-dessus
TEST: {{ci dessus}}                                                             ->> ci-dessus
TEST: {{par dessus}}                                                            ->> par-dessus
TEST: {{au delà}}                                                               ->> au-delà
TEST: {{ci devant}}.                                                            ->> ci-devant
TEST: Il faisait froid cet {{hiver la}}.                                        ->> hiver-là
TEST: {{Ceux la}} vont à la pêche.                                              ->> Ceux-là